Vous êtes sur la page 1sur 240

Nuclear Development

Externalities and Energy Policy:


The Life Cycle Analysis Approach

Workshop Proceedings
Paris, France
15-16 November 2001

NUCLEAR ENERGY AGENCY


ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT
ORGANISATION DE COOPÉRATION ET DE DÉVELOPPEMENT ÉCONOMIQUES
En vertu de l’article 1er de la Convention signée le 14 décembre 1960, à Paris, et entrée en vigueur le 30
septembre 1961, l’Organisation de coopération et de développement économiques (OCDE) a pour objectif de
promouvoir des politiques visant :
− à réaliser la plus forte expansion de l’économie et de l’emploi et une progression du niveau de vie
dans les pays Membres, tout en maintenant la stabilité financière, et à contribuer ainsi au
développement de l’économie mondiale ;
− à contribuer à une saine expansion économique dans les pays Membres, ainsi que les pays non
membres, en voie de développement économique ;
− à contribuer à l’expansion du commerce mondial sur une base multilatérale et non discriminatoire
conformément aux obligations internationales.
Les pays Membres originaires de l’OCDE sont : l’Allemagne, l’Autriche, la Belgique, le Canada, le
Danemark, l’Espagne, les États-Unis, la France, la Grèce, l’Irlande, l’Islande, l’Italie, le Luxembourg, la
Norvège, les Pays-Bas, le Portugal, le Royaume-Uni, la Suède, la Suisse et la Turquie. Les pays suivants sont
ultérieurement devenus Membres par adhésion aux dates indiquées ci-après : le Japon (28 avril 1964), la
Finlande (28 janvier 1969), l’Australie (7 juin 1971), la Nouvelle-Zélande (29 mai 1973), le Mexique (18 mai
1994), la République tchèque (21 décembre 1995), la Hongrie (7 mai 1996), la Pologne (22 novembre 1996),
la Corée (12 décembre 1996) et la République slovaque (14 décembre 2000). La Commission des
Communautés européennes participe aux travaux de l’OCDE (article 13 de la Convention de l’OCDE).

L’AGENCE DE L’OCDE POUR L’ÉNERGIE NUCLÉAIRE


L’Agence de l’OCDE pour l’énergie nucléaire (AEN) a été créée le 1er février 1958 sous le nom
d’Agence européenne pour l’énergie nucléaire de l’OECE. Elle a pris sa dénomination actuelle le 20 avril
1972, lorsque le Japon est devenu son premier pays Membre de plein exercice non européen. L’Agence
compte actuellement 27 pays Membres de l’OCDE : l’Allemagne, l’Australie, l’Autriche, la Belgique, le
Canada, le Danemark, l’Espagne, les États-Unis, la Finlande, la France, la Grèce, la Hongrie, l’Irlande,
l’Islande, l’Italie, le Japon, le Luxembourg, le Mexique, la Norvège, les Pays-Bas, le Portugal, la République
de Corée, la République tchèque, le Royaume-Uni, la Suède, la Suisse et la Turquie. La Commission des
Communautés européennes participe également à ses travaux.
La mission de l’AEN est :
− d’aider ses pays Membres à maintenir et à approfondir, par l’intermédiaire de la coopération
internationale, les bases scientifiques, technologiques et juridiques indispensables à une utilisation
sûre, respectueuse de l’environnement et économique de l’énergie nucléaire à des fins pacifiques ; et
− de fournir des évaluations faisant autorité et de dégager des convergences de vues sur des questions
importantes qui serviront aux gouvernements à définir leur politique nucléaire, et contribueront aux
analyses plus générales des politiques réalisées par l’OCDE concernant des aspects tels que l’énergie
et le développement durable.
Les domaines de compétence de l’AEN comprennent la sûreté nucléaire et le régime des autorisations, la
gestion des déchets radioactifs, la radioprotection, les sciences nucléaires, les aspects économiques et
technologiques du cycle du combustible, le droit et la responsabilité nucléaires et l’information du public. La
Banque de données de l’AEN procure aux pays participants des services scientifiques concernant les données
nucléaires et les programmes de calcul.
Pour ces activités, ainsi que pour d’autres travaux connexes, l’AEN collabore étroitement avec l’Agence
internationale de l’énergie atomique à Vienne, avec laquelle un Accord de coopération est en vigueur, ainsi
qu’avec d’autres organisations internationales opérant dans le domaine de l’énergie nucléaire.

© OCDE 2002
Les permissions de reproduction partielle à usage non commercial ou destinée à une formation doivent être
adressées au Centre français d’exploitation du droit de copie (CFC), 20, rue des Grands-Augustins, 75006
Paris, France. Tél. (33-1) 44 07 47 70. Fax (33-1) 46 34 67 19, pour tous les pays à l’exception des États-Unis.
Aux États-Unis, l’autorisation doit être obtenue du Copyright Clearance Center, Service Client, (508)750-
8400, 222 Rosewood Drive, Danvers, MA 01923 USA, ou CCC Online : http://www.copyright.com/. Toute
autre demande d’autorisation ou de traduction totale ou partielle de cette publication doit être adressée aux
Éditions de l’OCDE, 2, rue André-Pascal, 75775 Paris Cedex 16, France.
ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT

Pursuant to Article 1 of the Convention signed in Paris on 14th December 1960, and which came
into force on 30th September 1961, the Organisation for Economic Co-operation and Development (OECD)
shall promote policies designed:
− to achieve the highest sustainable economic growth and employment and a rising standard of
living in Member countries, while maintaining financial stability, and thus to contribute to
the development of the world economy;
− to contribute to sound economic expansion in Member as well as non-member countries in
the process of economic development; and
− to contribute to the expansion of world trade on a multilateral, non-discriminatory basis in
accordance with international obligations.
The original Member countries of the OECD are Austria, Belgium, Canada, Denmark, France,
Germany, Greece, Iceland, Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, Spain, Sweden,
Switzerland, Turkey, the United Kingdom and the United States. The following countries became Members
subsequently through accession at the dates indicated hereafter: Japan (28th April 1964), Finland (28th January
1969), Australia (7th June 1971), New Zealand (29th May 1973), Mexico (18th May 1994), the Czech
Republic (21st December 1995), Hungary (7th May 1996), Poland (22nd November 1996); Korea (12th
December 1996) and the Slovak Republic (14th December 2000). The Commission of the European
Communities takes part in the work of the OECD (Article 13 of the OECD Convention).

NUCLEAR ENERGY AGENCY


The OECD Nuclear Energy Agency (NEA) was established on 1st February 1958 under the name
of the OEEC European Nuclear Energy Agency. It received its present designation on 20th April 1972, when
Japan became its first non-European full Member. NEA membership today consists of 27 OECD Member
countries: Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece,
Hungary, Iceland, Ireland, Italy, Japan, Luxembourg, Mexico, the Netherlands, Norway, Portugal, Republic of
Korea, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States. The Commission of
the European Communities also takes part in the work of the Agency.
The mission of the NEA is:
− to assist its Member countries in maintaining and further developing, through international
co-operation, the scientific, technological and legal bases required for a safe, environmentally
friendly and economical use of nuclear energy for peaceful purposes, as well as
− to provide authoritative assessments and to forge common understandings on key issues, as
input to government decisions on nuclear energy policy and to broader OECD policy
analyses in areas such as energy and sustainable development.
Specific areas of competence of the NEA include safety and regulation of nuclear activities,
radioactive waste management, radiological protection, nuclear science, economic and technical analyses of
the nuclear fuel cycle, nuclear law and liability, and public information. The NEA Data Bank provides nuclear
data and computer program services for participating countries.
In these and related tasks, the NEA works in close collaboration with the International Atomic
Energy Agency in Vienna, with which it has a Co-operation Agreement, as well as with other international
organisations in the nuclear field.

© OECD 2002
Permission to reproduce a portion of this work for non-commercial purposes or classroom use should be
obtained through the Centre français d’exploitation du droit de copie (CCF), 20, rue des Grands-Augustins,
75006 Paris, France, Tel. (33-1) 44 07 47 70, Fax (33-1) 46 34 67 19, for every country except the United
States. In the United States permission should be obtained through the Copyright Clearance Center, Customer
Service, (508)750-8400, 222 Rosewood Drive, Danvers, MA 01923, USA, or CCC Online:
http://www.copyright.com/. All other applications for permission to reproduce or translate all or part of this
book should be made to OECD Publications, 2, rue André-Pascal, 75775 Paris Cedex 16, France.
INTERNATIONAL ENERGY AGENCY
9, RUE DE LA FÉDÉRATION, 75739 PARIS CEDEX 15, FRANCE

The International Energy Agency (IEA) is an autonomous body which was established in November 1974
within the framework of the Organisation for Economic Co-operation and Development (OECD) to
implement an international energy programme.

It carries out a comprehensive programme of energy co-operation among twenty six* of the OECD’s thirty
Member countries. The basic aims of the IEA are:

− To maintain and improve systems for coping with oil supply disruptions;
− To promote rational energy policies in a global context through co-operative relations with non-member
countries, industry and international organisations;
− To operate a permanent information system on the international oil market;
− To improve the world’s energy supply and demand structure by developing alternative energy sources
and increasing the efficiency of energy use;
− To assist in the integration of environmental and energy policies.

* IEA Member countries: Australia, Austria, Belgium, Canada, the Czech Republic, Denmark, Finland,
France, Germany, Greece, Hungary, Ireland, Italy, Japan, Korea, Luxembourg, the Netherlands, New
Zealand, Norway, Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom, the United States. The
European Commission also takes part in the work of the IEA.

ORGANISATION FOR ECONOMIC CO-OPERATION


AND DEVELOPMENT

Pursuant to Article 1 of the Convention signed in Paris on 14th December 1960, and which came into force on
30th September 1961, the Organisation for Economic Co-operation and Development (OECD) shall promote
policies designed:
− To achieve the highest sustainable economic growth and employment and a rising standard of living in
Member countries, while maintaining financial stability, and thus to contribute to the development of the
world economy;
− To contribute to sound economic expansion in Member as well as non-member countries in the process
of economic development; and
− To contribute to the expansion of world trade on a multilateral, non-discriminatory basis in accordance
with international obligations.
The original Member countries of the OECD are Austria, Belgium, Canada, Denmark, France, Germany,
Greece, Iceland, Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland,
Turkey, the United Kingdom and the United States. The following countries became Members subsequently
through accession at the dates indicated hereafter: Japan (28th April 1964), Finland (28th January 1969),
Australia (7th June 1971), New Zealand (29th May 1973), Mexico (18th May 1994), the Czech Republic (21st
December 1995), Hungary (7th May 1996), Poland (22nd November 1996); Korea (12th December 1996) and
the Slovak Republic (14th December 2000). The Commission of the European Communities takes part in the
work of the OECD (Article 13 of the OECD Convention).

© OECD 2002
Permission to reproduce a portion of this work for non-commercial purposes or classroom use should be
obtained through the Centre français d’exploitation du droit de copie (CCF), 20, rue des Grands-Augustins,
75006 Paris, France, Tel. (33-1) 44 07 47 70, Fax (33-1) 46 34 67 19, for every country except the United
States. In the United States permission should be obtained through the Copyright Clearance Center, Customer
Service, (508)750-8400, 222 Rosewood Drive, Danvers, MA 01923, USA, or CCC Online:
http://www.copyright.com/. All other applications for permission to reproduce or translate all or part of this
book should be made to OECD Publications, 2, rue André-Pascal, 75775 Paris Cedex 16, France.
FOREWORD

Improving the sustainability of energy systems by incorporating energy


externalities into related policy making represents a key challenge that has been
studied by both the International Energy Agency (IEA) and the OECD Nuclear
Energy Agency (NEA). The life cycle analysis (LCA) approach, which consists
of an evaluation of the potential environmental impacts of a product through its
life cycle “from cradle to grave”, provides a conceptual framework for a detailed
and comprehensive, comparative evaluation of energy supply options. LCA
could help national energy policy making by pointing to opportunities to
improve the sustainability of the full fuel cycle operations in OECD countries
(for example, by improving the sustainability of mining practices) and providing
quantitative input to the political debate on improving the sustainability of
energy systems.

For these reasons, the IEA and the NEA jointly organised a workshop held
in Paris on 15-16 November 2001 on the subject “Energy Policy and
Externalities: The Life Cycle Analysis Approach”. The workshop brought
together internationally recognised experts in this field with energy policy
makers to examine the state of the art in assessing and internalising external
costs of energy in power generation and transportation using LCA, as well as the
potential and limitations of the LCA method for its use in energy policy making.

These proceedings provide a record of the presentations made by the LCA


experts along with a summary of the discussions held among experts, policy
makers and members of the industry. The results of the expert analyses made
available during the workshop suggest that when monetised, energy externalities
can form a significant part of the total economic cost of certain sources of power
generation or transportation fuels. While large uncertainties remain, and may
prevent their direct application in policy making, the LCA results may at least
provide a qualitative guide to policy makers on the relative environmental
impacts of the different energy technologies and therefore on the implied
benefits of greater utilisation of low-impact technologies. They may also serve
as a basis for a constructive dialogue with key stakeholders on resource
development and use, recognising that different stakeholders have different
interests and responsibilities.

5
It is hoped that these proceedings of “Energy Policy and Externalities: The
Life Cycle Analysis Approach” will play a similar role, helping to inform the
debate among energy policy makers, the industry and other stakeholders on the
challenges and opportunities of improving the sustainability of energy use.

6
TABLE OF CONTENTS

Foreword..................................................................................................... 5

Executive Summary................................................................................... 9

Opening Session Energy Policy and Externalities: an Overview on


External Cost Issues

Energy Policy and Externalities: An Overview ..........23


David Pearce, Workshop Chairman

Session 1 Approaches & Issues


(Theory, Concepts, Definitions, Relevance to Decision
Making)

The ExternE Project: Methodology, Objectives and


Limitations................................................................................47
Ari Rabl, Joseph V. Spadaro

Session 2 External Costs of Energy/Electricity Life Cycles


(Results from Recent Authoritative Studies, Lessons
Learned, Uncertainties, Gaps)

A Life Cycle Perspective of Coal Use ............................. 65


Louis Wibberley
Well-to-wheel Energy Analysis Study............................. 79
Jean Cadu
From Life Cycle Analysis Approach to Monetarisation
of the Impacts: an Evaluation in Term of Decision Process .. 91
Marc Darras

7
Life Cycle Analysis as Basis for Evaluating
Environmental Impacts of Energy Production ............... 103
Edgar Furuholt

The External Cost of the Nuclear Fuel Cycle ................ 111


Caroline Schieber, Thierry Schneider
Hydropower – Internalised Costs and Externalised
Benefits .......................................................................... 131
Frans H. Koch
Life Cycle Assessment of Renewables: Present Issues,
Future Outlook and Implications for the Calculation
of External Costs ............................................................ 141
Paolo Frankl

Session 3 Comparative Assessments in Electricity and Transportation

LCA/External Costs in Comparative Assessment of


Electricity Chains. Decision Support for Sustainable
Electricity Provision?....................................................... 163
Alfred Voss
Life-cycle Analysis and External Costs in Transportation..183
Mark A. Delucchi

Round Table How to Use Internalisation of Externalities in Policy Making?

Energy Policy and Externalities: the Life Cycle


Analysis Approach ......................................................... 195
Maria Rosa Virdis

List of participants ................................................................................... 235

8
EXECUTIVE SUMMARY

Introduction

The International Energy Agency (IEA) and the Nuclear Energy


Agency (NEA) organised jointly a Workshop entitled “Energy Policy and
Externalities: The Life Cycle Analysis Approach” to discuss issues related
to assessing external costs of energy using the life cycle analysis approach
and investigate the relevance of this process for decision making.

The meeting, held on 15-16 November 2001 at the IEA Headquarters,


was chaired by David Pearce and attended by some 75 experts from
governmental bodies and industries. Bill Ramsay and Luis Echávarri opened
the meeting by welcoming the participants.

David Pearce identified four issues in his opening keynote address:


consistency between Life Cycle Analysis (LCA) and economic theory
generally; uncertainties with respect to health-related externalities; uncertainties
and relevance of discounting costs related to global warming; and the empirical
underpinnings for “disaster aversion” in externality estimates.

He stressed the need to ensure economic consistency in assessing


external costs noting that an environmental impact was only an externality if
it was not fully compensated for. Occupational risks (which may already be
accounted for in higher wage rates) and resource depletion (which might be
reflected in market prices) are two areas where at least some of the costs
have been internalised. The fact that most of the externality literature points
to global warming and health effects as dominating energy externalities is
important in the light of the controversy surrounding “dose-response”
effects. While the usual academic conclusion that more research is needed
always seems frustrating to policy makers, it should be recognised that we
do not seem to know enough about the economic valuation of life risks to be
confident about the kinds of adders being produced in externality studies.
Likewise, estimates of global warming damage neglect adaptation responses
and do not reflect adequately the allocation of costs and benefits both
among rich and poor countries and in time. Also, we need to look far more

9
rigorously at the way in which discounting should be integrated into damage
estimates in particular for impacts/damage occurring in the long term.

“Disaster aversion” refers to an apparent higher economic value attached


to a large number of deaths in a single accident compared to an equal number
of deaths spread over in a large number of more frequent types of accident.
The ExternE study and others have looked at this matter and attempted to
develop aversion factors that would lead to higher risk estimates for nuclear
power. However, there is little empirical literature to support this.

Finally, one can see the need for additional “meta-analysis” of the
external cost literature in order to compare in a systematic way results from
different studies and analyse differences in order to explain why damage
estimates differ (assumptions, local conditions, technologies, methods, …).

Approaches & issues

The methodology and main conclusions of the ExternE project, carried


out for the European Commission, were outlined by Ari Rabl, of the École
des Mines, as one example of approach to external cost assessment. The
methodology uses a bottom up approach to estimate the impacts of different
emissions from different power generation and transportation fuel options
through inventory of each emission, estimate of its dispersion, examination
of the impact based on the dose response-relationship (impacts being
measured essentially in terms of years of life lost) and the economic
valuation of these impacts. The results generally show that the estimated
external costs are much higher for fossil fuels (highest for coal and oil than
gas), and much lower for nuclear and renewables (although photo-voltaic
stands out as a renewable energy technology with relatively high impact).
These results are subject to a large number of uncertainties and quantitative
estimates are probably no more reliable than a factor of 3. Uncertainties
arise not only from data limitations, but also from difficulties in quantifying
certain impacts (ecosystem), assumptions about future management of
waste and improvements in technology, and intergenerational
considerations.

External costs of energy/electricity life cycles

The work on life cycle analysis presented by Louis Wibberley, BHP


Sustainable Technology, covers a range of energy sources and technologies
for steel and electricity production. A streamlined “cradle to gate” approach
was taken, with a focus on energy/technology comparisons, detailed

10
understanding of the production chain, and an attempt to look particularly
for opportunities to improve environmental performance as measured by
greenhouse gas emissions. The results show that the use of coal and natural
gas for steel-making may have comparable emission impacts, if slag from
the blast furnace is fully utilised as clinker for cement making, and off gases
from the furnace are used to produce electricity. They also indicate potential
gains from increased use of scrap inputs, and from using coal bed methane
at the mine. Overall a reduction in greenhouse gas intensity of up to 50%
could be obtained. In the power production sector, LCA shows the largest
possibilities for improvement through use of more efficient technologies,
use of biomass to displace coal and utilisation of fly ash in cement making.
One interesting technological possibility is combining solar thermal
technology with coal power generation, which improves net solar efficiency
to 30-40% (compared to 13% for photo-voltaic). Estimated additional costs
for large-scale use of solar thermal in an existing coal plant are about
0.04 cents US/kWh.

The results of “Well to Wheel Analysis” of transportation fuel


alternatives undertaken for a consortium of oil companies (Shell, BP, Exxon
Mobil) and GM by Argonne National Laboratory in the US were presented by
Jean Cadu, Shell. The analysis evaluates separately the fuel cycle (well to
tank) and the propulsion (and emission control) systems (tank to wheel). The
analysis assessed 75 different fuel pathways (gasoline diesel, Fischer Tropsch
diesel, ethanol, methanol, hydrogen, and natural gas) and 15 different
propulsion systems (conventional, conventional hybrid, and fuel cell) on a
single vehicle type (GM full-size pickup). The results show that in terms of
primary energy consumption, petroleum hybrid engines or fuel cell engines
are significantly better than most natural gas options and all renewable
options. However, in terms of greenhouse gas emissions, fuel-cell vehicles
offer a clear edge (although diesel-engine hybrid vehicles are close), but
renewable fuel options using ethanol (from biomass fermentation) offer the
lowest emissions. A similar study is now underway for Europe to reflect the
differences in oil supply, refining system, and vehicle efficiency (an Opel
Zafira is the vehicle model). The study will also examine a greater number of
renewable fuel pathways.

A critique of the life cycle analysis approach, was presented by


Edgar Furuholt, Statoil, following a brief introduction from Marc Darras,
Gaz de France. He noted that production chains often produce multiple
products, some for energy use, some for others and that allocation of the
emissions was to some degree arbitrary and must be handled with care.
Given the wide degree of variability of characteristics of oil and gas
production, any emission estimate could be derived given the appropriate
selection of wells, extraction processes etc. Similarly, impact assessments

11
fail to take into account unknown health and environmental impacts of new
chemicals. For example, in the case of the chemical MTBE, a life cycle
assessment was performed but failed to foresee possible health and
groundwater effects. The presentation argued that LCA has no objective
scale, contains too many assumptions, and is too complex to provide
transparent results. Therefore, it should not be used as the basis for
comparing widely different generating options or as the basis for
internalising external costs. On the other hand, it is valuable for systematic
descriptions of resource use and environmental impact characteristics, and
can be used more precisely when the production chains and technology
options are all very similar, or in choosing amongst locations for the same
technology option.

The work carried out by the “Centre d’étude sur l’évaluation de la


protection dans la domaine nucléaire” (CEPN) within the ExternE study,
was summarised by Caroline Schieber, CEPN, who presented its main
results with emphasis on factors affecting the estimates of nuclear power
external costs. Nuclear power impacts differ from the fossil-fuelled chain
impacts in that they are principally the result of exposure to radioactivity
from routine operation of fuel cycle facilities and power plants, and in the
case of accidents. According to the ExternE study, based upon the French
nuclear chain, most of the exposure results from the electricity generation
and fuel reprocessing phases and takes place over a very long period. The
external cost is largely attributed to impacts on workers while the cost of
impacts on the public are rather small (about 0.00002 per kWh). This
figure is not greatly increased by accidents, using the “large consensus”
assumption that such accidents would occur at a frequency of 1 per 100 000
reactor-years and that, in such an accident, 1% of the radioactive materials
would be released to the environment. Even if a risk aversion effect is
assumed, the figure for accidents would only be around 0.0001 per kWh,
still a small figure. Long-term impacts of mining waste are also quite low
and their management according to current regulations greatly reduces
releases as compared to past practices according to the data from
UNSCEAR reports.

Externalities from hydropower projects were addressed by Frans Koch,


IEA Implementing Agreement for Hydropower. He emphasised that the
motto: “avoid (environmental externalities), mitigate (damages that can’t be
avoided), compensate (damages that can’t be mitigated)”, adopted within
hydropower projects already actively contributes to reduce externalities.
The Implementing Agreement surveyed a large number of energy LCA
studies to better understand the relative position of hydropower. Emissions
of greenhouse gases from hydropower dams are normally quite low, with
few exceptions. The survey found that other positive benefits of hydropower

12
dams such as irrigation or flood control are not normally taken into account
by such studies. Energy security benefits are also not generally recognised,
and would represent a useful extension of LCA. Also not all environmental
impacts can be usefully internalised by an LCA (e.g. loss of visual amenity
is not usually included in LCA), and LCA does not take into account
cultural differences of the value of different amenities.

LCA analyses for photo-voltaic (PV) and wind were summarised by


Paolo Frankl, Ecobilancio Italia. He noted that there were a large range of
LCA values for PV in the literature. Estimates of LCA impacts, originally
quite high although they included only PV modules, have fallen quite
steeply despite including the balance of system materials in the assessment,
because of the decreasing requirements for inputs into solar PV modules
and the improved outputs. Whereas systems installed in the late 1990s could
be expected to require eight years before their energy output exceeded the
energy used to produce the panel, this figure could easily fall to 2.3 years
based on latest technology available. There are further possibilities to
reduce this through heat recovery. For amorphous silicon technologies, it is
already less than one year. Emissions from current PV are only about a
factor of 2.6 lower than emissions from the electricity displaced from the
Italian grid, but can easily be expected to improve to a factor of 20 in the
near future (and even higher factors may be possible with heat recovery).
For wind, externalities are quite low, although the operation phase produces
both noise and loss of visual amenity. Wind damage estimates are the
lowest of all the ExternE fuel cycles studied. The experience with PV shows
the need to look at LCA in a dynamic way, particularly with respect to new
technologies. A new international research project, ECLIPSE, will look at
the life cycle inventories for future power generation technologies,
focussing on PV, wind, fuel cells, biomass and CHP technologies.
Sensitivity analysis will look at the impact of rapid technological
improvement and differences in local conditions. The output, by the end of
2003, will be a report and guidelines on how to assess these technologies in
life cycle analysis work.

Comparative assessments in electricity and transportation

The results of LCA for power generation based on German conditions,


presented by Alfred Voss, Institute of Energy Economics and the Rational
Use of Energy, University of Stuttgart, show that coal (particularly lignite)
power generation have the highest external costs in terms of years of life
lost, followed by PV (because of its high energy intensity and low insolation
in Germany) and natural gas with nuclear, wind and hydro giving the lowest
results. In terms of costs, this implies coal/lignite have external costs around

13
3 cents per kWh, gas and PV around 1 cent, with nuclear wind and
hydro about 0.1 cents. If these external cost estimates are combined with
direct costs, nuclear, which is already nearly competitive with coal and
cheaper than natural gas, becomes the lowest cost option for power
generation. There are however, many uncertainties in terms of data, choices
of discount rate etc. LCA can provide valuable support to decision-makers
with regard to technology evaluation, comparison of future energy supply
options, cost benefit analysis of policy measures and extension of green-
accounting frameworks. Since uncertainties are still so large, it is better to
use LCA only where appropriate regionally differentiated pollutant-specific
damage estimates for cost internalisation are available (e.g. for SOx, NOx
and particulate matter). The risk of low probability accidents can be
internalised into the monetary accounting system by introducing liability
insurance obligations.

An analysis on upstream and life cycle emissions, including greenhouse


impacts, in the transport sector was presented by Mark Delucchi, University
of California Davis. The assessment of greenhouse gas impacts in the US
shows that against a baseline gasoline vehicle, the impact of including the full
fuel cycle generally reduces the relative advantages of alternative
transportation fuels. While a switch to diesel is estimated to save 30% as
compared to gasoline, the savings from natural gas/LPG are (around 20%),
for ethanol from corn (8%) and for battery electricity vehicles using power
from coal (6%) are much smaller. This is largely due to the use of LCA rather
than end-use comparisons. However, the results also show that there would
be large savings from the use of ethanol from fuel cells using methanol (39%)
or natural gas (50%), while ethanol from wood in a conventional engine
appears to have the greatest savings (63%). In external costs of motor vehicle
use, analysis results were presented for both air pollution and energy security
impacts (including SPR, military expenditures, macroeconomic costs and
pecuniary costs) as well as water pollution, noise and congestion impacts. The
results suggest that externalities amount to 1.2 US cents per mile travelled in
gasoline powered vehicle. The most significant externality is related to air
pollution. Costs associated with US defence, the SPR, and climate change are
quite insignificant. The only other variable of significance is the impact on the
economy, through the transfer of wealth outside the US (referred to as
“pecuniary externality”) and the oil price shock impacts on the economy. A
comparison of external costs and subsidies for different transportation modes
in the US (gas or electric cars, transit bus, light rail, heavy rail) showed that
subsidies available to public transit system greatly outweigh the benefit in
reduced externalities avoided. In the comparison of social costs of
transportation alternatives, differences in external cost, while not trivial, are
outweighed by the differences in direct costs or in subsidies.

14
Round table

A background paper prepared by the Secretariat was presented by


Maria Rosa Virdis, IEA, to initiate the roundtable discussion. First, she
remarked that previous speakers had stressed the limitations of LCA,
particularly for purposes of policy making (internalisation of externalities,
green taxation, standards) rather than its usefulness as a tool. She illustrated
through a couple of examples how LCA could be used to correct existing
policies or to set benchmarks for new policies. The first example relates to
estimated external costs of diesel versus gasoline vehicle use. Based on results
of the ExternE study, she suggested that health impacts (and damages) due to
particulate emissions from diesel vehicles in a sample of European cities seem
to be much higher than those caused by cars using gasoline and equipped with
catalytic converters. However, the damages, while approximately covered by
existing gasoline taxes are only fractionally covered in the case of diesel fuel
(with prevailing tax burdens on diesel and gasoline in Europe). In fact,
European countries on average tax diesel 20-40% less than gasoline, thus
favouring the former notwithstanding its higher health costs. Several policy
approaches were suggested to eliminate this distortion and to better internalise
external costs. The second example considered the proposed EC policy to
allow for a maximum subsidy of 5 cents/kWh for electricity from renewable
sources. This value was calculated on the basis of the external costs avoided
by renewables, again based on ExternE results. Implications for the relative
competitiveness of technologies and fuels were discussed, as well as some of
the advantages and limitations of subsidy policies. While indicating some of
the most obvious limitations of the LCA approach and possible areas of
improvement, she recommended further development of the methodology and
wider dissemination and use of its results as a tool to support policy making.

Following this presentation, five pannelists presented brief remarks on


the usefulness of life cycle analysis for government policy making and
business decision making.

Birgit Bodlund (Vattenfall) recounted the Swedish power industry


experience with cradle to grave/gate life cycle analysis. She cited three
values for LCA: informing political debate (the politicians were surprised
by the results and asked why they had not been informed earlier);
contributing to continuous improvement in design and operation of the
power system; and certifying the environmental qualities of the power
through Environmental Product Declaration (EPD 14025) in response to the
requests of customers. She noted that the good marks nuclear gets through
such analysis are not sufficient to answer criticism of nuclear power. The
complexity of LCA means that its implications are more readily taken up by
future decision-makers than by the current generation.

15
Marc Darras (Gaz de France) noted that decision-makers never limit
themselves to a strict cost-benefit analysis. LCA helps identifying the
various criteria, economic, social and environmental, that are involved in
making decisions and the associated uncertainties. It thus can be a valuable
tool in defining the context for complex decisions. However, such decisions
ultimately require the actors concerned to rely on their own judgement for
ensuring that appropriate weight is given to each criterion. LCA alone
cannot set external cost because of the uncertainties in the data, the
intergenerational dimension, and the exclusion of non-energy aspects.
Decision-makers also need to keep in mind the geographic distribution of
impacts, costs and benefits while LCA does not reflect it.

Ture Hammar (Danish Energy Agency) discussed how LCA evolved in


Denmark from more integrated approaches to pollution reduction beginning
with resource depletion and environmental impact assessment in the 1970s, to
a concerted effort to develop “clean” energy technologies in the 1980s to a
move to “internalising externalities” by eliminating subsidies for “dirty”
technologies and introducing environmental taxes and subsidies for “clean”
technologies. The present approach is also more comprehensive and
integrates sustainability. While LCA can be a useful tool in thinking globally
about energy matters in a sustainable development perspective, it cannot be
the basis for political decisions. For example, the Danish perspective on the
Swedish nuclear reactor near Copenhagen inevitably takes other factors into
account. LCA needs to try and capture the dynamics of technology
development. Another application of LCA is in support of “ecolabelling”,
which aims at leading the energy consumers towards choosing
environmentally sustainable production. The Danish electricity utilities have
introduced “ecolabelling” for electricity and heat and published results
comparing waste generation from different energy sources and found
generally that CHP is less polluting than conventional fossil generation.

Ron Knapp (World Coal Institute) stated that the coal industry supports
internalisation of all positive and negative externalities into energy costs but
not focusing on a single externality. He commented that 7 of the 10 largest
coal exporters and most coal importers are countries which do not have
limits on their GHG emissions under the Kyoto Protocol. He commented
that politicians are beginning to awaken to issues about energy security
despite the recent fall in oil prices but doubted whether the market, with the
“tunnel vision of bankers” was considering resource depletion in its
decisions. He stressed that while coal improved its environmental
performance dramatically over the past two generations, this change has
come about through political will. Government policy should not close off
technology options, but allow them to be improved e.g. CO2 capture and
storage technologies. When using LCA analysis, political boundaries and

16
implications for different countries should be take into account. For
example, two power stations in Tokyo, one coal-fired, one natural gas-fired,
have about the same amount of emissions on a life cycle basis but gas has
far lower emissions on an end-use basis and may thus be preferred over coal
by the user country. Power companies cannot be expected to make choices
on a life cycle basis unless there is an incentive from government policy
makers for them to do so.

Marcella Pavan (Electricity and Gas Regulatory Authority of Italy)


discussed the impact of market liberalisation on the role that regulators could
play in dealing with externalities. The regulatory authority is principally
concerned with promoting competition and ensuring high quality of service in
electricity and gas. Among its tasks are to review contracts and satisfy itself
regarding certain technical issues with respect to security of supply. In these
conditions, it is difficult to see how much economic textbook principles on
externalities can actually be applied in the regulatory context, given the
different government institutions that are involved at different stages of the
life cycle. Some Italian examples illustrate this point. First example, there are
taxes on sulphur and nitrogen oxide emissions, but the tax levels (53 and
104.8 per tonne, respectively) are less than 1% of the values of external
costs found by ExternE. A second area is the use of externality adders in the
valuation of DSM programmes. The law permits utilities to engage in energy
conservation programmes and allows cost recovery to include a component
that recognises reduced environmental damage. A third example is in the area
of carbon taxes, which not only differ for different fuels but exclude/exempt
autoproduction and ignore the effects of methane leakage of gas pipelines.
She stated that the real problem with externality internalisation or life cycle
analysis is the political will to use the results. This means more research needs
to be done on country-specific internalisation tools to take into account the
different policy contexts in each country. It must also take into account
market liberalisation, through the use of market-based instruments such as
tradable green or emissions certificates.

The debate between panellists and participants that followed the above
presentations focused on the usefulness of LCA and its limitations. The
discussions also identified some areas for future research. The role of taxes
and subsidies was addressed to a certain extent, with emphasis on the
challenge raised by uncertainties on external costs which limit their
relevance in fixing taxes but are not an excuse for political inaction in front
of polluting activities and potentially harmful subsidies.

LCA was found useful as a qualitative guide to policy-makers on the


relative environmental impacts of different energy technologies and
therefore on the implied benefits of greater utilisation of low-impact

17
technologies. The role of LCA in communicating about impacts of energy
and transport activities was underlined. For example, LCA results
synthesising a wide range of data and analyses may facilitate information
flows between government and industry policy makers and the public, and
serve as a basis for a constructive dialogue with key stakeholders on
resource development and use. LCA offers a framework to compare global
impacts (i.e. from greenhouse gas emissions) and site specific impacts. The
approach may be used for fine tuning existing technology choices,
identifying opportunities for improvement in these technologies and guiding
companies in choosing suppliers.

The limits to LCA applicability are due mainly to: uncertainties on


results; incomplete scope (some impacts are not covered); static analysis
(technology progress cannot be reflected); and site-specific character of the
method. Uncertainties on the results raise a challenge for using the results in
policy making. Since results are site-specific it is not easy to draw generic
conclusions from LCA studies. The aggregated nature of LCA, encom-
passing an entire chain of activities taking place in various jurisdictions
limits its relevance for policy making. Since the scope of LCA does not
cover security of supply, ecosystem integrity, biodiversity, or social
impacts, the approach is not comprehensive enough for measuring the
sustainability of an energy system. Furthermore, LCA focuses on what is
“in the light” (i.e. what can be analysed readily), but is not much help for
criteria which cannot be readily quantified and differences in social value
systems between countries are not reflected. Last but not least, technology
developments, which may significantly change life cycle impacts, are not
taken into account since the assessment is static and does not reflect
dynamic system evolution.

The possible areas for future research identified include: assessment of


externalities such as security and diversity of supply, as well as loss of
forest cover; further investigations in the field of discount rates applicable in
the very long term and value of statistical life; incorporation of dynamics,
technology progress, in LCA; evaluation of energy policy measures with
LCA; further effort to reduce uncertainties in ExternE; and establishment of
a data base containing information on externality assessment and the way it
is being used (possibly under the umbrella of the IEA and NEA).

Main findings from the workshop

David Pearce summarised the workshop through comments from an


environmental economist’s perspective. He suggested that the discussions
should distinguish between damage done (impact) by a system (energy or

18
transport in the present case) and a policy-relevant externality, i.e. cost
associated with a damage (or a benefit) that is not internalised in the price
paid by consumers. In the area of fuel taxes, although they are already very
high, it is questionable whether environmental externalities are already
incorporated. Similarly, the positive “employment externality” of energy
production may be irrelevant as most economic theory would suggest that in
near-full employment economies it does not exist. He questioned whether
the incorporation of “accident aversion” as an external cost factor had an
empirical basis. He wondered to what extent “September 11-type” events
are already incorporated in the analyses of ExternE. He argued that while
some questioned whether aesthetics or other qualitative externalities could
be valued, there is research, some if it already incorporated in LCA, that
does attempt to quantify such externalities through the willingness to pay to
avoid them. However, biological diversity impacts are much harder to
define since we do not know what they are, how to measure them, or how
they are affected by different power generation or transportation fuel
options. He stated that probably the most important issue in LCA is the
question of time and discounting which is particularly critical in discussing
the greenhouse gas emission problem since the damage caused by global
warming will occur mainly in a rather long term future and will vary with
time. He stressed the substantial interface between LCA and the economics
of resource depletion and added that the question of whether current
economic decisions reflect resource depletion is not yet answered. Finally,
he reminded the audience that “politics will decide” how and to what extent
externalities are ultimately incorporated into economic decisions and
politicians are not making the best of all possible decisions in the best of all
possible worlds.

19
Opening Session

ENERGY POLICY AND EXTERNALITIES:


AN OVERVIEW ON EXTERNAL COST ISSUES

Workshop Chairman:
Professor David Pearce, University College London

21
ENERGY POLICY AND EXTERNALITIES: AN OVERVIEW

David Pearce
University College London (United Kingdom)

1. The uses of externality adders

Substantial progress has been made in estimating the monetary value of the
environmental impacts of different energy systems. Perhaps the best known study
in Europe is that sponsored by the European Commission and known as the
ExternE programme [10-15,16-19]. In the USA a comparable project is that jointly
sponsored by the US Department of Energy and the European Commission [35,
36-37, 40, 42, 39, 38, 41]. There are many others. In each case what is sought is a
monetary value of an environmental impact arising from a unit of energy, usually
standardised as a kilowatt hour (kWh). These environmental impacts are usually
termed “externalities”. An externality exists if two conditions are met. First, some
negative (or positive) impact is generated by an economic activity and imposed on
third parties. Second, that impact must not be priced in the market place, i.e. if the
effect is negative, no compensation is paid by the generator of the externality to
the sufferer. If the effect is positive, the generator of the externality must not
appropriate the gains to the third party, e.g. via some price that is charged. In the
energy externality literature, the procedure of expressing the externalities in, say,
th
cents or milli-euros (1 000 of an Euro = m ) per kWh results in an “adder”. An
adder is simply the unit externality cost added to the standard resource cost of
energy. Thus, if an electricity source costs X m to produce or deliver, the final
social cost of it is (X+y) m where y is the externality adder.

While externality adders have been researched most in the context of energy,
they are increasingly being estimated for other economic sectors, notably transport
[16, 18] and agriculture [22,47,54]. What are the uses of such figures?

First, such figures could be used to guide investment decisions. If major


electricity sources remain in public or quasi-public ownership, then the full
social cost of electricity by different sources could be used to plan future
capacity, with preference being given to the source with the lowest social cost.
Where electricity is privately owned, then full social cost can be used by

23
regulators to guide new investment or to act as an effective environmental tax,
leaving the private owners to respond accordingly.

Second, adders can be used to estimate environmental taxes. While the use of
adder estimates is not typically used in this way, the UK has at least two taxes
based on externality estimates which, in turn, contain elements of external
estimates taken from energy adder studies (the aggregates tax and the landfill tax).

Third, adders may be used as an input into modified national accounts. Here
the idea is to replace GNP (or, more correct, net national product, NNP) with a
measure that accounts for the depreciation of natural resources, so that “green”
NNP becomes NNP – depreciation on resource – damage to the environment).

Fourth, adder estimates may be used for awareness raising, i.e. simply
drawing attention to the fact that all energy sources have externalities which give
rise to economically inefficient allocations of resources.

Fifth, adders might help with some notion of priority setting for
environmental policy. The basic principle would be that, as a first
approximation, attention should be paid to those activities generating the highest
externalities. Better still, activities should be prioritised by some cost-benefit
principle, so that adders are ranked according to the ratio of reduced adders to
the cost of securing that reduction.

Clearly, then, estimating externality adders is potentially very informative


for policy purposes. While a huge amount of research, and a large sum of
money, has gone into estimating adders, problems remain. In some cases, the
search for figures that can be used has perhaps obscured the need to think more
fundamentally about how adders are derived, the conditions under which they
might be considered reasonably valid, and the uses to which they are put. We
therefore focus on just a few of the more important issues.

2. Externality and environmental impact: consistency with economic


theory

Some of the adder literature proceeds as if any “external” impact constitutes


an externality. But this is not so. The second attribute of the definition given
previously is that any third party impact must be unpriced, i.e. uncompensated or
unappropriated. To illustrate the problem, consider occupational health effects, i.e.
impacts on those who work in the energy industries, and, for that matter in
industries that supply the energy industries, e.g. mining, or which dispose of waste
from the energy industries. Since the adder literature adopts a life cycle approach,
these impacts could be important. But one of the methodologies used to estimate

24
the value of risks to life is in fact based on the notion that wages in risky
occupations are higher, other things being equal, precisely because of the risks
involved. In other words, risks are “internalised”, i.e. compensated for, in wage
payments. If this is true, then one cannot include occupational risks in adder
estimates. Indeed, there is a contradiction in using “values of statistical life”, most
estimates for which come from wage-risk studies, whilst simultaneously arguing
that occupational risks reflect an externality.

We can illustrate by looking at the ExternE 1995 estimates of the


externalities from a pressurised water reactor (PWR) [14, p. 191]. Taking a
discount rate of 3%, the estimates there suggest that the total externality in
-2
m ECU (now m ) is 6.00 E-02 or 6 × 10 m ECU. Of this, 5.73 E-02 consists of
occupational effects. In other words, over 95% of the externality is accounted for
by occupational effects. But if these effects are internalised in wages, they
should not be included and the resulting externality is trivial at 0.27 E-02.

Occupational effects are perhaps one of the easier sources of double


counting to identify. But there are similar problems with accidents in the
transportation phases of the life cycle analysis. Some accidents are undoubtedly
truly external, but in many cases risks are already internalised in the decision to
drive or go by train etc. The conclusion has to be that more care needs to be
devoted to the consistency between the estimates and the underlying economic
theory that must be obeyed if the estimates are to be regarded as useful for
policy purposes.

3. Valuing statistical lives

A second major issue concerns the valuation of health effects in combined


LCA and valuation studies. A glance at both the externality adder literature and
cost-benefit studies of pollution controls shows that (a) health damages tend to
dominate measures of externality and health benefits dominate cost-benefit
studies, (b) within health effects, changes in life expectancy dominate. Table 1
shows a selection of studies relating to air pollutants and reveals that health
benefits account for a minimum of one-third and a maximum of nearly 100% of
overall benefits from pollution control. Moreover, in most cases these benefits
exceed the costs of control by considerable margins. Health benefits therefore
“drive” positive benefit-cost results. Nor is this outcome peculiar to the
European Union. The US EPA’s retrospective and prospective assessments of
the Clean Air Act produce extremely high benefit-cost ratios, e.g. 44 for the
central estimate of benefits and costs [58]. Moreover, EPA regards these as
probable underestimates. In turn, the benefits are dominated by health benefits
(99% if damage to children’s IQ is included). The EPA’s analysis has, however,

25
been subjected to very critical analysis [29,50]. By contrast, the European
studies appear not to have attracted much by way of critical comment.

It may be the case that there are very high benefit-cost ratios for air
pollution control, but there are at least two reasons for a feeling of unease about
the results that are being obtained:

(1) The relevant studies tend to omit ecosystem benefits, despite the fact
that, for acidifying substances in the wider Europe, ecosystem
protection is the driver for the UN ECE region air pollution Protocols
under the Convention on Long Range Transport of Air Pollution
(LRTAP). If the presumption of the Convention Parties that ecosystem
damage is of dominant importance is correct, this would suggest that
benefit cost ratios are substantially higher than the factors of three to
five being recorded in the European studies. Some would regard this is
adding to doubts about the analysis, rather than reducing them.
(2) The European studies suggest that benefits exceed costs even for
scenarios defined in terms of “maximum technologically feasible
reduction” (MFR) of pollutants, i.e. scenarios in which the most
pollutant-reducing technologies are used. Such scenarios should be
characterised by very high marginal abatement costs at very high levels
of pollution reduction, precisely the context where one would expect
incremental benefits to be less than incremental costs. While the benefit
cost ratio does appear to fall for such scenarios relative to other more
modest abatement targets, the reduction is not dramatic and benefits
continue to exceed costs. Thus, AEA Technology [6] finds a benefit cost
ratio of 2.17 for a MFR scenario, compared to 2.87 for practical targets
based on the relevant Protocol. The incremental benefit cost ratio of
going from Protocol targets to “MFR” targets is 1.6.

A similar picture emerges in the ExternE studies. Taking the UK National


Implementation study, but omitting occupational health for the reasons given
previously, Table 2 shows the percentage of externality due to “public health”
effects and to global warming. We return to the global warming issues shortly.
The estimates suggest the following conclusions. First, global warming and public
1
health effects account for virtually all the externality from all fuel cycles . Second,
for coal, oil and orimulsion the two effects are broadly comparable. For gas, global
warming is the overwhelming impact. For nuclear and the renewables, public
health dominates and global warming is relatively unimportant.

1. This needs to be qualified because the ExternE estimates, along with other studies,
generally omit ecosystem impacts beyond crop damage.

26
Table 1. Health benefits as a percentage of overall benefits
in recent cost-benefit studies

Study Title and subject area Benefits as %


total benefits

Holland and Krewitt [23] Benefits of an Acidification 86-94%. Total benefits


Strategy for the European cover health, crops and
Union: reductions of SOx, NOx, materials.
NH3 in the European Union.
AEA Technology [1] Cost Benefit Analysis of 80-93%. Total benefits
Proposals Under the UNECE cover health, crops,
Multi-effect Protocol: buildings, forests,
reductions of SOx, NOx, NH3, ecosystems, visibility.
VOCs.
AEA Technology [2]; Economic Evaluation of the 52-85% depending on
Krewitt et al.[28] Control of Acidification and inclusion or not of chronic
Ground Level Ozone: reductions health benefits. Total benefits
of NOx and VOCs. SO2 and include health, crops,
NH4 held constant. materials and visibility.
AEA Technology [3] Economic Evaluation of Air B/C ratio of 0.32 to 0.46
Quality targets for CO and for CO. Costs greatly
Benzene. exceed benefits for
benzene. Benefits consist
of health only.
AEA Technology [4] Economic Evaluation of B/C ratios of 3.6 to 5.9.
Proposals for Emission Ceilings Health benefits dominate.
for Atmospheric Pollutants.
AEA Technology [6] Cost Benefit Analysis for the VOSL + morbidity
Protocol to Abate Acidification, accounts for 94% of
Eutrophication and Ground benefits. B/C ratio = 2.9.
level Ozone in Europe.
IVM, NLUA and IIASA Economic Evaluation of Air 32-98%. Total benefits
[25]; Quality for Sulphur Dioxide, include health and
Olsthoorn et al.[43]. Nitrogen Dioxide, Fine and materials damage.
Suspended Particulate Matter
and Lead: reductions of these
pollutants.
Note to Table 1: we have selected results using VOSL (value of statistical life) rather
than “VOLY” (value of a life year) since the latter are not correctly estimated in the
studies that also provide VOLY results. See text for discussion.

27
Table 2. Percentage contribution of public health and
global warming damages in all damages: ExternE, UK

Coal Oil Orimulsion Gas Nuclear Wind Biomass


Health 44 48 41 20 81 68 85
Gwarming 53 50 56 79 15 22 9
TOTAL 97 98 97 99 96 90 93
Source: adapted from AEA Technology [5].
Notes: health refers to public health only.

The fact that health and global warming effects dominate is important since
both are very controversial both in terms of the “dose-response” literature and in
terms of economic valuation. We deal first with just a few of the health issues. A
fuller discussion can be found in [44].

The epidemiology literature linking air pollution to health effects is large.


In terms of risk of death, there are two types of literature. The first relates acute
episodes of pollution to life risks and the second, a far smaller literature, relates
chronic exposure to air pollution to life expectancy. Most of the economic
valuation literature deals with the former, i.e. with acute effects. But it is
becoming increasingly clear that the chromic exposure epidemiology is more
important, although acute studies still have a role to play. One of the problems
with acute studies is that they may tell us numbers of people dying from acute
effects but not the period of life that is foreshortened. There is a debate as to
whether the life periods concerned are very short indeed, a matter perhaps of just
a few days in OECD countries, or whether what evidence we do have on life
foreshortening understates the true effects. This debate is well rehearsed in the
contributions in [45]. The second problem is that, surprisingly, the epidemiology
tells us little about the age groups that are affected. But the available evidence
suggests, as one might expect, that the relevant deaths tend to occur in much
older age-groups, usually in the over 70s. Combing these two likely facts of
short duration loss of life expectancy with the age effect suggests that the
relevant economic value will be the willingness to pay of over 70s to avoid days
rather than years of life loss. Again surprisingly, we have limited evidence on
how willingness to pay relates to age, but what we have suggests that it will be
lower than the willingness to pay of median age groups involved in accidents.
Yet it is studies relating to the latter that tend to determine the “value of
statistical life” (VOSL) used in cost-benefit and life cycle externality studies.
Figures such as 3 million are common in the ExternE studies, for example. No-
one is suggesting that the values of the older generation do not matter, but we do
have to question whether values such as 3 million can possibly be relevant to
such impacts.

28
Turning to chronic mortality, while the relevant studies are far fewer and,
even then, some of them simply borrow dose-response coefficients from
previous studies, there is some suggestion that chronic exposure foreshortens life
by perhaps six months and maybe one year. Supposing this to be true, the
question then arises of what economic value we should attach to such
epidemiological effects. If the effect of chronic exposure is to induce illness
which may itself ultimately result in life foreshortening, then we should
definitely be concerned to estimate the willingness to pay to avoid that illness.
For the reduced life expectancy itself, however, it looks as if the correct value is
what we are willing to pay today to extend our lives by, say, six moths when we
are in our 70s or even 80s. We have very little evidence on what these sums are
since the willingness to pay studies we do have relate to risks that are reduced
(or increased) with effectively an immediate effect (e.g. a road accident).

Johannesson and Johansson [26] report a contingent valuation study in


Sweden where adults are asked their willingness to pay for a new medical
programme or technology that would extend expected lifetimes conditional on
having reached the age of 75. Respondents are told that on reaching 75 they can
expect to live for another 10 years. They are then asked their WTP to increase
lifetimes by 11 years beyond 75, i.e. the “value” of one extra year. The results
suggest average willingness to pay across the age groups of slightly less than
10 000 SEK using standard estimation procedures and 4 000 SEK using a more
conservative approach, or say 600-1 500. This is for one year of expected life
increase. Using the formula:

VOSL(a ) = VOLY ∑1 /(1 + r ) T −a


t

As suggested in [26] these values are consistent with “normal” VOSLs of


30 000 to 110 000, substantially less than the VOSLs being used in cost-
benefit and externality adder studies. Since T-a is obviously less the older the
age group, then the relevant VOSLs will decline with age.

It is perhaps worth noting that “rules of thumb” used in the ExternE work
are not valid. The approach to valuing a “life year” in the ExternE studies
proceeds as follows. A “value of a life year” or VOLY2 can be thought of as the
annuity which when discounted over the remaining life span of the individual at
risk would equal the estimate of VOSL.

2. The ExternE notation is “YOLL” for “year of life lost”.

29
Thus, if the VOSL of, say, £1.5 million relates to traffic accidents where the
mean age of those involved in fatal accidents is such that the average remaining
life expectancy would have been 40 years, then:

VOLY = VOSL/A
-n
where A = [1-(1+r) ]/r and n is years of expected life remaining and r is the
3
utility discount rate. Examples are shown below in Table 3 for n = 40 years.

Table 3. Deriving VOLYs from VOSLs: examples

VOSL m Utility discount rate Utility discount rate Utility discount rate
= 0.3%. A = 37.6 = 1.0%. A = 32.8 = 1.5%. A = 29.9
1.0 26 595 30 460 33 445
1.5 39 894 45 690 50 167
2.0 53 190 60 920 66 890
3.0 79 787 91 138 100 000

These VOLY numbers can then be used to produce a revised VOSL


allowing for age. At age 60, for example, suppose life expectancy is 15 years.
The VOSL(60) is then given by:
VOSL(60) = ΣVOLY/(1+r)
T-60

where T is life expectancy. In the case indicated, this would be, at a 1% discount
rate and a “standard” VOSL of 1 million:
VOSL(60) = (30 460).(13.87) = 422 480

The result is that the age-related VOSL declines with age and this appears
to accord with the findings noted earlier that willingness to pay probably does
decline with age. The generalised formula for age related VOSL is:
VOSL(a ) = [VOSL(n) / A]∑ 1 /(1 + r ) T − a
t

where a is the age of the individual or group at risk, T is life expectancy for that
group, VOSL(a) is the age-adjusted VOSL and VOSL(n) is the “normal” VOSL.

3. Another way of saying the same thing is that VOLY = VOSL/Discounted expected
lifetime. Strictly, the relationship holds only when utility of consumption is constant in
each time period.

30
There are several reasons for doubting the usefulness of the VOLY
approach when it is based on a VOSL.

First, the basis of the VOLY approach is the life-cycle consumption model
with uncertain lifetime. It is well known that such models assume utility depends
on consumption alone and not on the length of life. Lifetime utility does indeed
vary with life expectancy but the route is via consumption not via time itself. It
seems unlikely that individuals are indifferent to time remaining. There are also
additional restrictions on the model to ensure that WTP is proportional to the
discounted value of life expectancy. Thus, it can be questioned whether the
underlying theory needed to derive VOLYs from VOSLs is itself tenable.

Second, the theory forces the age-distribution of VOLYs to take on a


monotonically declining form: VOLY simply declines with age. What evidence
we have, however, suggests that willingness to pay follows an inverted “U”
shape curve, rising to a median age and then falling. If so, the VOLY construct
is a poor representation of “true” WTP over the lifetime of individuals.

What can we conclude on the health effects and valuation in externality


adder studies? While the usual academic conclusion that “more research is
needed” always seems frustrating to policy makers, the fact is that we do not
know enough about the epidemiology and we certainly do not know enough
about the economic valuation of life risks to be confident about the kinds of
adders being produced in externality studies. In some cases, being more certain
of the absolute magnitudes of the adders may not matter too much. For example
if we simply wish to prioritise investments by social cost, a ranking may not be
affected by what values we use. But if we wish to use the values to set, say,
energy taxes, then the absolute magnitudes do matter.

4. Global warming

Table 2 above showed that the other major component of externality values
derives from global warming effects. Everyone is familiar with the scientific
uncertainties in warming studies. To those uncertainties we must add economic
valuation uncertainties. Uncertainty is not a reason for neglecting economic
valuation – there is a widespread but erroneous view that if we avoid trying to
estimate economic values what we will end up with is a more certain base for
policy making than if we do not. The reality is that whatever decisions we make
about global warming policies they will all imply some set of economic values.
It is better to be as explicit as we can about the numbers rather than masking
them by procedures that allegedly do not use them.

31
Table 4 shows some of the results taken from economic studies of global
warming. The relevant magnitude is the economic value of one tonne of
greenhouse gas emitted now. This must allow for residence time in the atmosphere
and the fact that greenhouse gases are cumulative. The relevant concept is
therefore a discounted economic value of damage due to the “marginal” (i.e. extra)
tonne of pollutant. This is the basis of Table 4.

Table 4 shows that the relevant values per tonne carbon vary substantially
with the discount rate assumed. This is hardly surprising. Unfortunately, while
economists are reasonably good at estimating social discount rates for a single
nation, the relevant discount rate for the world as a whole is a more elusive
concept. Yet it is the relevant one because the damages recorded in Table 4
relate to the world as a whole. More complex still, economists have not yet
secured a consensus on what the relevant rate would be for very long-lived
environmental effects of the kind that would typify global warming damages
and, just as relevant for externality adders, nuclear waste disposal. The most
promising contribution to date appears to be that of Weitzman [59] who shows
that the long term discount rate should almost certainly be declining with time
(Annex 1 to this paper is an attempt to derive Weitzman’s result in a much
simpler way). Our first observation about global warming estimates, then, is that
we need a far more rigorous look at the way in which discounting should be
integrated into the damage estimates.

A second observation is that all of the estimates in Table 4 are based on the
“dumb farmer” syndrome. They do not make any allowance for adaptation to
global warming. Yet, if we know anything at all, we know that people do not
stand idly by and do nothing in the face of environmental damage.
Unfortunately, we appear to have only one set of studies that give us any idea
what would happen to the damage estimates if we do assume adaptation. In an
important contribution, a volume edited by Mendelsohn and Neumann [31]
shows that total damages to the US economy could be zero instead or positive
once adaptation is assumed.

There are problems with this claim if we wish to extrapolate it to the global
damage estimates underpinning the marginal damage estimates in Table 4. First,
the estimates relate to the market sectors of the US economy only. Yet it is the
non-market sectors such as ecosystem functioning that perhaps give the greatest
cause for concern. Second, adaptation in the USA is likely to be greater than in
the developing world where fewer technological options are available.
Nonetheless, the global warming damage estimates in the externality adder
literature do need revisiting in light of the Mendelsohn-Neumann findings.

32
Table 4. Marginal damages from greenhouse gases ($ tonne carbon)

Study Estimate $ tC . Base year prices: 2000


Period 1991-2000 2001-2010 2011-2020 2021-2030
Nordhaus [32] 9.3 9.3 – –
Nordhaus [33]
p = 0.03, best guess 6.8 8.7 11.0 12.8
p = 0.03, expected value 15.4 23.0 33.9 –
Nordhaus and Boyer [34] 6.4 9.1 11.9 15.0
Fankhauser [21]
with p =0,0.005,0.03 26.0 29.2 32.4 35.6
with p =0 62.5 – – 80.5
with p =0.03 7.0 – – 10.6
Cline, [9]
with s = 0 7.4-158.7 9.7-197.1 12.5-238.1 15.1-282.9
Peck and Teisberg [46]
with p =0.03 12.8-15.4 15.4-17.9 17.9-23.0 23.0-28.2
Maddison [30] 7.6-7.8 10.4-10.8 14.2-14.8 18.8-19.4
Eyre et al. [20]
with s = 0 181.8 190.7
with s = 1 93.4 92.2
with s = 3 29.4 25.6
with s = 5 11.5 9.0
with s = 10 2.6 1.3
Tol [57] 14.1 16.6 19.2 23.0
Roughgarden and Schneider [48]:
lower bound = Nordhaus, upper 6.4-14.1 7.7-16.6 10.2-20.5 12.8-26.9
bound = Tol
Notes:
• s = social discount rate. Eyre et al. [20], estimates are for 1995-2004 and 2005-2014 and the
estimates here exclude equity weighting. p = utility discount rate and s = the overall discount
rate. Roughgarden and Schneider’s [48] ranges derive from placing the models of Fankhauser
[21], Cline [8], Titus [55] and Tol [57] into Nordhaus’s DICE model framework [33]. The upper
end of the range should, strictly, coincide with the marginal damage estimates in Tol [57].
• Most original estimates are in 1990$ and we have assumed an escalation of 2.5% p.a. inflation.
Note also that Table 1 shows the considerable sensitivity of estimates to discount rates. The
discount rates given for Fankhauser’s estimates relate to the pure time preference rate
component, p, only. According to Fankhauser [21] his social cost estimates based on the
distribution of values for p are equivalent to a “best guess” value of 0.5% for p. To this must be
added a value for the elasticity of the marginal utility of income multiplied by the expected
growth rate. Fankhauser and Eyre et al. [21, 20], take the elasticity to be about unity, so that the
only variable is the expected long term economic growth rate of income per capita. Setting this
at 1.6-1.8% pa, the discount rate would be 2.2 to 2.3%. Accordingly, the values in
Eyre et al. [20] of 0-3% are more relevant for purposes of comparison.

33
While there are many problems with the global warming estimates, one that
is much discussed and debated deserves some comment. In the ExternE work the
global warming damages are “equity weighted”. Global warming damages are
spread across the world and affect both rich and poor countries. But poor
countries have substantially lower incomes than rich countries so that one Euro’s
damage to them has a higher “disutility” value than one Euro’s damage to rich
countries. Cost-benefit analysis typically works with notions of willingness to
pay that do not reflect this adjustment for different utilities of a money unit. But
there is no unique way to do cost-benefit so it is perfectly legitimate to seek to
maximise utility-adjusted benefits and costs. This is what the ExternE
programme does [17]. In terms of the economic value of a tonne of carbon, the
effect of equity weighting is to raise the value per tonne. This is because
damages done to richer individuals tend to be used as the numeraire, with the
result that damages to poorer individuals are increased relative to what they
would have been without equity weighting. As shown in Eyre et al. [20], the
effect of equity weighting is approximately to double the marginal damage
estimates. Once again, if the policy issue of concern is one of prioritising fuel
cycles the use of equity weighting may not matter too much. But if absolute
levels of damages matter, then it is crucial to justify equity weighting.
Unfortunately, studies using equity weighting are not very forthcoming on what
this justification is. The ExternE reports suggest that it is consistent with
maximising “utility” as opposed to willingness to pay-based measures of costs
and benefits. This is correct. But several questions then arise. First, why do we
seek to maximise equity-weighted net benefits in the global warming context but
not in any other context? Second, if we do adopt equity weights, what is the
justification for selecting one particular set of weights rather than another?

As far as the first question is concerned, the ExternE reports suggest that
equity, and by implication equity weighting, is integral to the Framework
Convention on Climate Change (FCCC) [20, p. 9]. This is debatable. The equity
notions in the FCCC relate primarily to the concept of differentiated
responsibility, i.e. since rich countries are bigger emitters of greenhouse gases
they bear more responsibility and should act first. It is a substantial leap from
this idea to one of weighting costs and benefits. More telling is the fact that
equity weighting in cost-benefit analysis of global warming control will raise the
benefit-cost ratio of taking action. In other words, more global warming control
would be undertaken with equity weighting than without it. This seems fair until
we recall that whatever is spent on global warming control is not spent on other
things, and the other things may include foreign aid, technology transfer etc., all
of which benefits the poorer countries. In turn this suggests that we should either
equity weight all policy measures that have an effect on poor countries or we
should not equity weight any of them. Isolating global warming and ring-fencing
it as if it is unique is not a tenable proposition. The situation could be even more
complex, since global warming expenditures are likely to benefit the

34
descendants of the current poor rather than the poor now. Yet the descendants
are likely to be richer than the current poor, so that the policy of weighting
damages may simply reinforce a tendency to divert resources from solving the
problems of the current poor. This was pointed out by Schelling [49].

Suppose, however, that we do accept equity weights. What are the relevant
weights? ExternE [20] suggests that the weights reflect diminishing marginal
utility but they actually select a specific value for the weighting procedure.
Table 5 shows what is being implied for different values of the elasticity of the
marginal utility of income function and for different ratios of income. Table 5
shows incomes differing by a factor of 2 and a factor of 20 (the latter is the ratio
of GNP per capita in high income to low and medium income countries).

Table 5. Equity weighting examples

Elasticity of the marginal utility of -0.5 -1.0 -2.0 -5.0


income = e
Social value of high income Y1 relative 0.7 0.5 0.25 0.03
to low income Y2 if Y1/Y2 = 2
Social value of Y1 if Y1/Y2 = 20 0.2 0.05 Neg Neg
-e
Note: the relevant formula is w = (Y1/Y2) where w is the weight on Y1.

What the Eyre et al.[20], and ExternE reports do is to select e = -1.0. For an
income differential of 2 this would imply that we value the higher income
group’s marginal income as being just 50% of the value to the low income
group. This seems potentially fair. But the income differential is actually 20, not
2, so that choosing -1.0 gives a weight of only 5% to income gains in the rich
country. Many people might think this is also fair, but it is categorically not how
we behave. If it was it would be impossible to explain why OECD countries
devote far less than one per cent (on average) of their GDP on foreign aid and
very much more than this on domestic life saving programmes.

The point here is not to assert that equity weighting is wrong. Indeed, not
adopting equity weights amounts to choosing the equity weights implicit in the
prevailing distribution of income. There is no escape from equity weighting! The
issue is whether the studies adopting specific estimates for externality adders
should adopt equity weights without (a) explaining why a specific set of weights
is chosen, and (b) explaining why those weights are relevant to global warming
but not to anything else.

35
5. “Disaster aversion”

A final issue that is very relevant to the externality adders literature is the
treatment of disaster aversion. The idea here is that lives are at risk from energy
sources. The usual procedure in the life cycle studies is to estimate accident rates
and then value the resulting accidents at the relevant VOSL. In this case, use of
“standard” VOSLs is probably correct because those at risk are the general
population. It is widely thought, however, that the population is not indifferent
between, say, 10 deaths in one accident and 10 deaths in 10 accidents each with
one death. This is the notion of “disaster aversion” whereby the economic value
attached to the former event would be higher than for the equivalent number of
deaths in the ten accidents. The issue is obviously particularly relevant for
nuclear power externalities, but it is also relevant for, say, gas explosions
affecting the general public. The ExternE adder estimates do not in fact contain
disaster aversion factors, but the ExternE background papers have discussed the
issue quite extensively. While great ingenuity has been brought to bear on the
kinds of aversion functions that might be specified, remarkably little empirical
work appears to have been done to test whether people really are averse to
disasters.

Ball and Floyd [7] reviewed studies of disaster aversion and concluded that
“there is very little evidence for differential risk aversion by the public where
this is based upon number of fatalities”. For example, Jones-Lee and Loomes
[27] found that, in a transport context, the risk of large-scale accidents did not
contribute to public willingness to pay for safety improvements. Ball and Floyd
[7] also cited a study, Hubert et al. [24], which assessed the disaster aversion of
senior managers in the petroleum, chemical and transport industries, as well as
elected officials, and found that disaster aversion was significant in this group of
people.

However, Slovic et al. [51], found that, in the nuclear context, the perceived
risk was much greater than the actual risk, despite a perception of having the
lowest annual number of fatalities compared to the other risk contexts studied.
This discrepancy was assigned to the perceived potential for disaster. Slovic et
al. [51,52] found that accidents also sent signals about the possibility and nature
of future accidents, and asserted that, for nuclear accidents, these secondary
impacts may be most important in this because the public perception of nuclear
accidents is of poorly understood risks with potentially catastrophic impacts. A
core damage accident may send ominous signals that the technology is out of
control, even if the number of injuries or deaths was small, and that could be
very damaging to the nuclear industry as a whole.

As it stands then, there is an urgent need to test for disaster aversion. It


could have a substantial effect on externality adders for nuclear power and

36
perhaps for natural gas. So far, however, we have little evidence to suggest that
people are averse to collective deaths in the manner suggested by some of the
theoretical literature.

6. Conclusions

Many other issues in the externality adder literature could have been
addressed. Probably the biggest omission here is the extent to which we are
justified in “borrowing” figures from other studies and using them in studies
such as those by ExternE. This is the issue of “benefits transfer” and it is very
much debated in the environmental economics literature. Major omissions in the
adder studies relate to the absence of “meta-studies” even though some of the
adder studies refer to meta-analysis. But it seems that what is usually meant is
that the literature on damages has been surveyed. Proper meta-analysis involves
statistical efforts to explain the variance in damage estimates and comparatively
few of these exist.

What we can say is that, thanks to the substantial efforts of exercises such
as ExternE, we are far better informed about externality adders than we were a
few years ago. This is to be welcomed. The problem remains that the theoretical
underpinnings are, in some cases, still weak.

37
Annex 1

Why the long run discount rate declines with time

The discount rate, r, needs to be distinguished from the discount factor,


t
1/(1+r) . It is the discount factor that gives the weight applied to each time
period. Suppose the discount rate and hence the discount factor is not known
with certainty and is a random variable. Suppose it takes the values 1…6% each
with a probability of 0.167. Table A1 shows the relevant values.
Table A1. Values of the discount factor

r DF10 DF50 DF100 DF200


1 0.9053 0.6080 0.3697 0.1376
2 0.8203 0.3715 0.1380 0.0191
3 0.7441 0.2281 0.0520 0.0027
4 0.6756 0.1407 0.0198 0.0004
5 0.6139 0.0872 0.0076 0.0000
6 0.5584 0.0543 0.0029 0.0000
Sum 4.1376 1.4898 0.5900 0.1589
Sum/6 0.7196 0.2483 0.0983 0.0265
r* 3.34% 2.82% 2.34% 1.83%
Note: DF10 = discount factor for year 10, etc. r* is the value of r that solves the equation
shown in the text.

While the weighted average (expected value) of the discount rate stays the
same in all periods (3.5%), the discount factor obviously varies with time. The
value of the implicit discount rate, r*, is given by the equation:

1
=
∑ DF t ,i
………i = n
(1 + r*) t n
where n = the number of possible discount rates, DF is the discount factor and t
is time.

Table A1 shows that the implicit discount rate goes down over time even though
the average discount rate stays the same for each period.

38
REFERENCES

[1] AEA Technology, (1998a), Cost-benefit Analysis of Proposals Under the


UNECE Multi-pollutant Multi-effect Protocol, Report to UK Department
of Environment, Transport and Regions, London and to UNECE Task
Force on Economic Aspects of Abatement Strategies, Geneva.

[2] AEA Technology, (1998b), Economic Evaluation of the Control of


Acidification and Ground Level Ozone, Report to DGXI of the European
Commission, Brussels.

[3] AEA Technology, (1998c), Economic Evaluation of Air Quality Targets


for CO and Benzene, DGXI, European Commission, Brussels.

[4] AEA Technology, (1998d), Economic Evaluation of Proposals for


Emission Ceilings for Atmospheric Pollutants. Report to DGXI, European
Commission, Brussels

[5] AEA Technology, (1998e), Power Generation and the Environment – a


UK Perspective, Report to the European Commission, DGXII, Brussels.

[6] AEA Technology, (1999), Cost-benefit Analysis for the Protocol to Abate
Acidification, Eutrophication and Ground Level Ozone in Europe,
Ministry of Housing, Spatial Planning and the Environment, The Hague,
Netherlands, Publication No. 133.

[7] D.J. Ball, P.J. Floyd, (1998), Societal Risks, London: Health & Safety
Executive.

[8] W. Cline, (1992), The Economics of Global Warming, Institute for


International Economics, Washington DC

[9] W. Cline, (1993), Modelling Economically Efficient Abatement of


Greenhouse Gases, Paper presented to United Nations Conference on
Global Environment, Energy and Economic Development, Tokyo.

39
[10] European Commission, (1995a), ExternE: Externalities of Energy.
Volume 1: Summary, Brussels (Belgium).

[11] European Commission, (1995b), ExternE: Externalities of Energy.


Volume 2: Methodology, EUR 16521 EN, Office for Official Publication
of the European Communities, Luxembourg.

[12] European Commission, (1995c), ExternE: Externalities of Energy.


Volume 3: Coal and Lignite, EUR 16522 EN, Office for Official
Publication of the European Communities, Luxembourg.

[13] European Commission, (1995d), ExternE: Externalities of Energy.


Volume 4: Oil and Gas, EUR 16523 EN, Office for Official Publication of
the European Communities, Luxembourg.

[14] European Commission, (1995e), ExternE: Externalities of Energy.


Volume 5: Nuclear, EUR 16524 EN, Office for Official Publication of the
European Communities, Luxembourg.

[15] European Commission, (1995f), ExternE: Externalities of Energy. Volume


6: Wind and Hydro, EUR 16525 EN, Office for Official Publication of the
European Communities, Luxembourg.

[16] European Commission, (1998a), ExternE: Externalities of Energy.


Volume 7: Methodology Update, EUR 19083, Office for Official
Publication of the European Communities, Luxembourg, 1999.

[17] European Commission, (1998b), ExternE: Externalities of Energy.


Volume 8: Global Warming, EUR 18836, Office for Official Publication
of the European Communities, Luxembourg, 1999.

[18] European Commission, (1998c), ExternE: Externalities of Energy.


Volume 9: Fuel Cycles for Emerging and End Use Technologies,
Transport and Waste, EUR 18887, Office for Official Publication of the
European Communities, Luxembourg, 1999.

[19] European Commission, (1998d), ExternE: Externalities of Energy.


Volume 10: National Implementation, EUR 18528 Office for Official
Publication of the European Communities, Luxembourg, 1999.

[20] N. Eyre, T. Downing, R. Hoekstra, K. Rennings, R.Tol, (1997), Global


Warming Damages, Final Report of the ExternE Global Warming Sub-
Task, DGXII, European Commission, Brussels.

40
[21] S. Fankhauser, (1995), Valuing Climate Change: the Economics of the
Greenhouse, Earthscan, London.

[22] O. Hartridge, D.W. Pearce, (2001), Is UK Agriculture Sustainable?


Environmentally Adjusted Economic Accounts For UK Agriculture.
Economics, University College London. Available at: www.cserge.ucl.ac.uk.

[23] M. Holland, W. Krewitt, (1996), Benefits of an Acidification Strategy for


the European Union, European Commission, DGXI, Brussels.

[24] P. Hubert et al., (1991), Elicitation of decision-makers’ preferences for


management of major hazards, Risk Analysis, 11:2, 199–206.

[25] IVM, NILU and IIASA, (1997), Economic Evaluation of Air Quality for
Sulphur Dioxide, Nitrogen Dioxide, Fine and Suspended Particulate
Matter and Lead, Report to DGXI, European Commission, Brussels.

[26] M. Johannesson, P-O. Johansson, (1996), To Be or Not to Be, That is The


Question: An Empirical Study Of the WTP For An Increased Life Expectancy
at an Advanced Age, Journal of Risk and Uncertainty, 13, 163-174.

[27] M. Jones-Lee, G. Loomes, (1995), Scale and Context Effects in the Valuation
of Transport Safety, Journal of Risk and Uncertainty, 11, 183–203.

[28] W. Krewitt, M. Holland, A. Trukenmüller, T. Heck and R. Friedrich, (1999),


Comparing Costs and Benefits of Strategies to Combat Acidification and
Ozone in Europe, Environmental Economics and Policy Studies, 2, 249-266.

[29] R. Lutter, (1998), An Analysis of the Use of EPA’s Benefit Estimates in


OMB’s Draft Report on the Costs and Benefits of Regulation, Comment
98-2, AEI-Brookings Center for Regulatory Studies, Washington DC.

[30] D. Maddison, (1994), The Shadow Price of Greenhouse Gases and


Aerosols, CSERGE, University College London, mimeo.

[31] R. Mendelsohn, J. Neumann (eds), (1999), The Impact of Climate Change


on the US Economy, Cambridge University Press, Cambridge.

[32] W. Nordhaus, (1991), To Slow or Not to Slow: the Economics of the


Greenhouse Effect, Economic Journal, 101, 407, 920-937.

[33] W. Nordhaus, (1994), Managing the Global Commons: the Economics of


Climate Change, MIT Press, Cambridge, Mass.

41
[34] W. Nordhaus and J.Boyer, (2000), Warming the World: Economic Models
of Global Warming, The MIT Press, Cambridge, Mass.

[35] Oak Ridge National Laboratory and Resources for the Future, (1992), US-
EC Fuel Cycle Study: Background Document to the Approach and Issues,
Oak Ridge: Oak Ridge National Laboratory. (Counted as Report 1 in the 8
volume series of reports).

[36] Oak Ridge National laboratory and Resources for the Future, (1994a),
Estimating Fuel Cycle Externalities: Analytical Methods and Issues.
Report 2, New York: McGraw Hill.

[37] Oak Ridge National Laboratory and Resources for the Future, (1994b),
Estimating Externalities of Coal Fuel Cycles, Report 3, New York:
McGraw Hill.

[38] Oak Ridge National Laboratory and Resources for the Future, (1998a),
Estimating Externalities of Natural Gas Fuel Cycles, Report 4, New York:
McGraw Hill.

[39] Oak Ridge National Laboratory and Resources for the Future, (1996),
Estimating Externalities of Oil Fuel Cycles, Report 5, New York:
McGraw Hill

[40] Oak Ridge National Laboratory and Resources for the Future, (1994c),
Estimating Externalities of Hydro Fuel Cycles, Report 6, New York:
McGraw Hill.

[41] Oak Ridge National Laboratory and Resources for the Future, (1998b),
Estimating Externalities of Biomass Fuel Cycles, Report 7, New York:
McGraw Hill.

[42] Oak Ridge National Laboratory and Resources for the Future, (1995),
Estimating the Externalities of Nuclear Fuel Cycles, Report 8, New York:
McGraw Hill.

[43] X. Olsthoorn, M. Amann, A. Vartonova, J. Clench-Aas, J. Cofala,


K. Dorland, C. Guerreiro, J. Henriksen, H. Jansen, S. Larssen, (1999), Cost-
benefit Analysis of European Air Quality Targets for Sulphur Dioxide,
Nitrogen Dioxide and Fine and Suspended Particulates in Cities,
Environmental and Resource Economics, 14, 333-351.

42
[44] D.W. Pearce, (2000), Valuing Risks to Life and Health: Towards
Consistent Transfer Estimate in the European Union and Accession
States, Report to DGXI, European Commission, Brusssels, Available at
www.cserge.ucl.ac.uk.

[45] D.W. Pearce, C. Palmer (eds.), (2002), Economics and Epidemiology: The
Nature and Economic Value of Health Effects from Air Pollution,
Cheltenham: Edward Elgar, forthcoming, Preliminary papers can be
viewed at: www.unece.org/env/nebei.

[46] S. Peck, T. Teisberg, (1993), Global Warming Uncertainties and the


Value of Information: An Analysis Using CETA, Resource and Energy
Economics, 15, 1, 71-97.

[47] J. Pretty, C. Brett, D. Gee, R. Hine, C. Mason, J. Morison, H. Raven,


M. Rayment, G. van der Bijl, (2000), An Assessment of the Total External
Costs of UK Agriculture, Agricultural Systems, 65(2), 113-136.

[48] T. Roughgarden, S. Schneider, (1999), Climate Change Policy: Quantifying


Uncertainties for Damages and Optimal Carbon Taxes, Energy Policy, 27,
415-429.

[49] T. Schelling, (1992), Some Economics of Global Warming, American


Economic Review, 82(1), 1-14.

[50] H. Sieg, V.K. Smith, H.S. Banzaf, R. Walsh, (2000), Estimating the
General Equilibrium Benefits of Large Policy Changes: the Clean Air Act
Revisited, National Bureau of Economic Research, Working Paper 7744,
NBER, Cambridge, Mass.

[51] P. Slovic, B. Fischhoff, S. Lichtenstein, (1980), Facts and Fears:


Understanding Perceived Risk in Societal Risk Assessment, in R. Schwing
and W. Albers (eds.), Societal Risk Assessment: How Safe is Safe
Enough? Plenum, New York, 181-216.

[52] P. Slovic, B. Fischhoff, S. Lichtenstein, (1982), Risk Aversion, Social


Values and Nuclear Safety Goals, Journal of the Transactions of the
American Nuclear Society, 41, 448–9.

[53] P. Slovic, B. Fischhoff, S. Lichtenstein, (1984), Modelling the Societal


Impact of Fatal Accidents, Management Science, 30:4, 464–474.

43
[54] R. Steiner, L. McLaughlin, P. Faeth, R. Janke, (1995), Incorporating
Externality Costs into Productivity Measures: a Case Study Using US
Agriculture, in V. Barbett, R. Payne, R. Steiner (eds.), Agricultural
Sustainability: Environmental and Statistical Considerations, New York:
Wiley, 209-30.

[55] J. Titus, (1992), The Costs of Climate Change to the United States, in
S. Majumdar (ed.), Global Climate Change: Implications, Challenges and
Mitigation Measures, Pennsylvania Academy of Science, Easton, Pa.

[56] R. Tol, (1995), The Damage Costs of Climate Change: Towards More
Comprehensive Estimates, Environmental and Resource Economics, 5, 353-
374.

[57] R. Tol, (1999), The Marginal Costs of Greenhouse Gas Emissions, The
Energy Journal, 20,1, 61-81.

[58] US Environmental Protection Agency, (1997), The Benefits and Costs of


the Clean Air Act 1970-1990, USEPA, Washington DC.

[59] M. Weitzman, (2000), Just keep discounting, but…, in P. Portney,


J. Weyant (eds.), Discounting and Intergenerational Equity, Washington
DC.: Resources for the Future, 23-30.

44
Session 1

APPROACHES & ISSUES


(THEORY, CONCEPTS, DEFINITIONS,
RELEVANCE TO DECISION MAKING)

45
THE EXTERNE PROJECT:
METHODOLOGY, OBJECTIVES AND LIMITATIONS

1
Ari Rabl, 2Joseph V. Spadaro
1
École des Mines, Paris (France)
2
International Atomic Energy Agency, Vienna (Austria)

Abstract

This paper presents a summary of recent studies on external costs of energy


systems, in particular the ExternE (External Costs of Energy) Project of the
European Commission. To evaluate the impact and damage cost of a pollutant,
one needs to carry out an impact pathway analysis; this involves the calculation of
increased pollutant concentrations in all affected regions due to an incremental
3
emission (e.g. µg/m of particles, using models of atmospheric dispersion and
chemistry), followed by the calculation of physical impacts (e.g. number of cases
of asthma due to these particles, using a dose-response function). The entire so-
called fuel chain (or fuel cycle) is evaluated and compared on the basis of
delivered end use energy. Even though the uncertainties are large, the results
provide substantial evidence that the classical air pollutants (particles, NOx and
SOx) from the combustion of fossil fuels impose a heavy toll, in addition to the
cost of global warming. The external costs are especially large for coal; even for
“good current technology” they may be comparable to the price of electricity. For
natural gas the external costs are about a third to a half of coal. The external costs
of nuclear are small compared to the price of electricity (at most a few %), and so
are the external costs of most renewable energy systems.

47
1. Introduction

In recent years there has been much progress in the analysis of public health
risks of energy systems, thanks to several major projects to evaluate the external
costs of energy in the USA [8,15] and in Europe [3,4]. Of these, the ExternE
(External Costs of Energy) Project of the European Commission has the widest
scope and it is being continually updated to incorporate the latest scientific findings
(external costs are damage costs imposed on others without compensation). This
paper presents key results for damage costs and mortality risks due to the routine
operation of the major energy technologies, as per ExternE [4].

To keep the presentation simple, we skip over much of the details and we
show approximate results that are typical rather than precise for a specific site.
We focus on mortality because it accounts for most of the damages (about 85%
according to ExternE [4]) when they are weighted in monetary terms. Note that
the damage cost estimates in the various publications are not always the same
because the methodology has been evolving. In particular, there is a profound
difference before and after 1996, due to the publication of a major study, by a
team at Harvard University [9], of the cumulative impacts of air pollution on
mortality, and due to a change in the valuation of air pollution mortality (loss of
life expectancy, rather than number of deaths). The latter change is appropriate
to permit meaningful comparisons between different causes of death with very
different loss of life per death, in particular accidents and air pollution. Our
paper is similar to Krewitt et al. [7], but places more emphasis on variability
with site and technology, and on comparisons of nuclear with other fuel chains
and with risks of everyday life.

2. Methodology

To evaluate the impact and damage cost of a pollutant, one needs to carry
out an impact pathway analysis, tracing the passage of the pollutant from the
place where it is emitted to the affected population. The principal steps of this
analysis can be grouped as follows:

• Emission: specification of the relevant technologies and the environmental


burdens they impose (e.g. kg of NOx per TWh emitted by power plant).

• Dispersion: calculation of increased pollutant concentrations in all affected


regions (e.g. incremental concentration of ozone, using models of
atmospheric dispersion and chemistry for ozone formation due to NOx).

48
• Impact: calculation of the dose from the increased exposure and
calculation of impacts (damage in physical units) from this dose, using a
dose-response function (e.g. number of cases of asthma due to this
increase in ozone).

• Cost: the economic valuation of these impacts (e.g. multiplication by the


cost of a case of asthma).

The impacts and costs are summed over all receptors of concern. For air
pollutants (other than globally dispersing greenhouse gases) from fossil fuels the
geographic range extends over thousands of km. The entire so-called fuel chain (or
fuel cycle) is evaluated and compared on the basis of delivered end use energy.
For the fossil fuel chains most of the damage cost arises from combustion in the
power plant; for nuclear and for renewable energy, by contrast, the impacts
upstream and/or downstream from the power plant are very important. The key
assumptions of ExternE [4] are summarised in Table 1.

Since the results, other than for globally dispersing greenhouse gases, can
vary strongly with the site where a pollutant is emitted, there is the question how
to generalize site-specific results to representative numbers that may be needed for
decision making. ExternE [4] has evaluated over fifty sites in different countries of
the EU, and the results are sufficiently representative to permit general
conclusions to be drawn [18,17]. Here we show results for central Europe, with an
2
average (land and water) population density of 80 persons/km ; for other regions
the impacts of PM10 (particulate matter with <10 m diameter), SO2 and NOx need
to be scaled in proportion to the population density within a radius of a thousand
km around the pollution source.

The goal of the monetary valuation of damages is to account for all costs,
market and non-market. For example, the valuation of a hospitalization should
include not only the cost of the medical treatment but also the willingness to pay
to avoid the suffering. If the willingness to pay for a non-market good has been
determined correctly, it is like a price, consistent with prices paid for market
goods. Economists have developed several tools for determining non-market
costs. Of these tools contingent valuation has enjoyed increasing popularity in
recent years, and the results of well-conducted studies are considered sufficiently
reliable.

49
Table 1. Key assumptions for the calculations in this paper [4]

Atmospheric dispersion models


Local range: ISC gaussian plume model [1].
Regional range (Europe): Harwell Trajectory Model [2] as implemented in
ECOSENSE software of ExternE.
Ozone impacts based on EMEP model [16], as
interpreted by Rabl & Eyre [11].
Global warming Physical impacts according to IPCC [6].
Impacts on health
Form of dose-response Linearity of incremental impact with an
a)
functions incremental dose for all health impacts.
Mortality Dose-response function slope fDR = 4.1 E-4
YOLL (years of life lost) per person per year per
3
g/m derived from increase in all-cause age-
specific mortality due to PM2.5 [9], by
integrating over age distribution.
b)
Nitrate and sulfate aerosols Dose response functions for nitrates same as for
PM10.
Dose response functions for sulfates same as for
PM2.5 (slope = 1.7 times slope of PM10 functions).
Radionuclides Linear dose-response functions:
0.05 fatal cancers/man·Sv,
0.12 nonfatal cancers/man·Sv,
0.01 severe effects hereditary /man·Sv.
Micropollutants Only cancers have been quantified (As, Cd, Cr,
Ni and dioxins).
Impacts on plants Dose-response functions for crop loss due to
SO2 and ozone.
Impacts on buildings and Corrosion and erosion due to SO2 and soiling
materials due to particles.
Impacts not quantified but Reduced visibility due to air pollution;
potentially significant disposal of residues from fossil fuels.
Economic valuation
Valuation of premature death Proportional to reduction of life expectancy,
with value of a YOLL (years of life lost) derived
from VSL = 3.1 M :
v
YOLL = 0.083 M per YOLL (year of life lost)
Valuation of cancers 0.45 M nonfatal cancers,
1.5 to 2.5 M (depending on YOLL) fatal cancers,
1.5 M average for cancers from chemical
carcinogens.
Discount rate 3% unless otherwise stated;
Costs for nuclear are shown for 0% “effective
discount rate” (=discount rate – escalation rate of
cost).
a) For air pollutants the terms concentration-response or exposure-response function have also been used.
b) Nitrates and sulfates are secondary pollutants created from NOx and SO2 emissions, respectively, by
chemical reactions in the atmosphere.

50
It turns out that damage costs of air pollution are dominated by mortality.
The key parameter is the so-called value of statistical life VSL – an unfortunate
term for what is really the willingness to pay for reducing the risk of premature
death. In ExternE [4], a European-wide value of 3.1 M was chosen for VSL,
close to similar studies in the USA. Unlike previous studies which simply
multiplied the number of premature deaths by VSL, ExternE [4] bases the
valuation on the years of life lost (YOLL). Since there had been no studies that
determine the value of a life year directly, ExternE [4] calculated the value of a
YOLL on theoretical grounds by considering VSL as the net present value of a
series of discounted annual values. The ratio of VSL and the value of a YOLL
thus obtained depends on discount rate and latency; it is typically in the range of
20 to 40. Here the value of a YOLL due to air pollution is taken as 0.083 M .

3. Damage cost per kg of pollutant

Before proceeding to fuel chain results, we present the damage costs per kg
of pollutant for typical European conditions in Table 2. Some indication of the
variability with site and stack conditions is given in the notes under the table.

The uncertainties are large. Rabl & Spadaro [13] have analysed them in
terms of lognormal distributions which are appropriate because the calculation
of damage costs is essentially multiplicative. The uncertainties can be expressed
as geometric standard deviation g, interpreted in terms of multiplicative
confidence intervals of the lognormal distri

bution: if a cost has been estimated to be g (geometric mean § median)


with geometric standard deviation g, the probability is approximately 68% that
2 2
the true value is in the interval [ g/ g, g. g] and 95% that it is in [ g/ g , g. g ].
For the damage costs of PM10, NOx and SO2, Rabl and Spadaro [13] estimate that
g is in the range of 3 to 4.

The largest sources of uncertainty lie in the epidemiology and in the “value of
statistical life” VSL. While ExternE [4] assumes VSL = 3.1 M , there is no
general consensus and one could argue for other values in the range of 1 to 5 M .
There is also considerable uncertainty about the relation between VSL (which has
been determined for accidents) and the value of a YOLL due to air pollution,
because it involves the period at the end of life about which valuation studies are
only just beginning.

The uncertainty due to the monetary valuation of a YOLL can be avoided


by comparing the risks of technologies directly on the basis of YOLL per kWh.
For this purpose we convert the /kg numbers for PM10, NOx and SO2 to

51
YOLL/kg by assuming as an approximate relation that 85% of the respective
damage costs is due to mortality, valued at 83 000 /YOLL. We do not indicate
the YOLL due to global warming because that information is difficult to extract
from the numbers in ExternE [4]. We note, however, that a large part of the
damage cost estimate of 0.029 /kg reflects loss of life due to the expected
increase in tropical disease.

Table 2. Typical damage costs per kg of pollutant emitted


by power plants in Europe [18,17]

a)
Pollutant Impact Cost , /kg
PM10 (primary) mortality and morbidity 15.4
SO2 (primary) crops, materials 0.3
SO2 (primary) mortality and morbidity 0.3
SO2 (via sulfates) mortality and morbidity 9.95
NO2 (primary) mortality and morbidity small
NO2 (via nitrates) mortality and morbidity 14.5
NO2 (via O3) crops 0.35
NO2 (via O3) mortality and morbidity 1.15
VOC (via O3) crops 0.2
VOC (via O3) mortality and morbidity 0.7
CO (primary) morbidity 0.002
As (primary) cancer 171
Cd (primary) cancer 20.9
Cr (primary) cancer 140
Ni (primary) cancer 2.87
1.85 × 10
7
Dioxins, TEQ cancer
CO2 Global warming 0.029
a) Variation with site and stack conditions (stack height, exhaust temperature, exhaust velocity):
• No variation for CO2.
• Weak variation for dioxin (non-inhalation pathways): factor of § 0.7 to 1.5.
• Weak variation for secondary pollutants: factor of §0.5 to 2.0.
• Strong variation for primary pollutants: factor of §WRIRUVLWH§to 3 for stack
conditions (up to 15 for ground level emissions in big city).

4. Results for fuel chains

Multiplying the cost per kg by the emission rates, Table 3, one readily finds
the cost per kWh. However, we warn against the temptation to cite cost/kWh or

52
YOLL/kWh numbers out of context, for instance “the health risks of coal”.
Quite apart from the variation of impacts with the site of an installation, the very
term “fuel chain” is misleading, because it suggests a simple monolithic system
while the reality is a complex chain whose elements can consist of a variety of
different processes and technologies at different sites, emitting very different
rates of pollution. Furthermore, thanks to ever more stringent environmental
regulations there has been a continual reduction of specific emissions (a factor of
3 to 10 during the past decade, except for CO2 for which the reductions have
been smaller).

Table 3. Emission of air pollutants and resulting years of life lost


YOLL/TWh for fossil power plants, for typical European conditions

Emission, g/kWh YOLL/TWh

PM10 SO2 NOx PM10 SO2 NOx Total


Coal, current 0.15 6 3 23.3 601.8 439.9 1 065
Coal, new 0.06 0.30 0.50 9.3 30.1 73.3 113
Oil, current 0.15 6 1.4 23.3 601.8 205.3 830
Oil, new 0.07 0.40 0.60 10.9 40.1 88.0 139
Gas, current negligible Smalla) 1.1 0.0 0.0 161.3 161
Gas, new negligible Smalla) 0.2 0.0 0.0 29.3 29
a) SO2 emission depends on composition of natural gas; in most cases it is negligible.
Note: “Current” = with typical emissions of existing fossil plants in USA and France
in 1995; “new” = with estimated emissions of large new power plants built in USA
since January 2000.

To illustrate this point we list in Table 3 the measured emissions of PM10,


SO2 and NOx, for typical fossil plants in USA and France during the nineties, as
well as estimated emissions for large new plants built in the EU after January
2000. For fossil fuel chains we show only emissions and YOLL due to the power
plant, rather than for the entire fuel chain, because they account for almost the
entire health impact; the YOLL due to emission of toxic metals is negligible, the
emitted quantities being extremely small. For the renewables, by contrast, the
health impacts arise mostly or exclusively from the production of the materials
for the power plant. Results are plotted in Figure 1. The number for nuclear is
based on the current technology of France, including reprocessing [3]; it
comprises the impacts over the entire globe and a time horizon of 100 000 years.
Assessments for other countries have found roughly comparable results [8,15,4].

Damage costs are plotted in Figure 2, showing separately the costs due to
the classical air pollutants (PM10, NOx and SO2), due to global warming

53
(including upstream emissions of CO2 and CH4, expressed as CO2 equivalent),
and due to radiation. It is interesting to note that the damage costs for existing oil
and coal fired power plants are very large, larger than the production costs of
electricity, about 25 to 50 m /kWh. The damage costs for nuclear correspond to
zero effective discount rate (= discount rate – escalation rate of costs). Because
of the uncertainties and controversies surrounding intergenerational discounting,
we show the impacts and costs separately for the near (before 100 yr) and far
(after 100 yr) future.

Figure 1. Comparison of public health risks in terms of YOLL/TWh,


for fuel chains in the EU, with emissions of Table 3

Coal, 1990s technology 1065

2000 technology 113

Oil, 1990s technology 830

2000 technology 139

Gas, 1990s technology 161

2000 technology 29

uclear, far future > 100 years 8

near future < 100 years 1

Biomass, High 100

Low 20

Photvoltaics, High 11

Low 4

Wind, High 9 YOLL per TWh


Low 4

Note: “High” and “low” for renewables indicate typical range of estimates of ExternE
[4]. This graph includes only risks due to PM10, SO2, NOx and radiation. For nuclear
only a single technology is shown (French, with reprocessing), but YOLL are separated
into near (before 100 yr) and far (after 100 yr) future.

For nuclear the numbers cover all stages of the fuel cycle, and include
waste disposal and major accidents – although any estimate of the latter items is
controversial. All of the damage cost of low-level radiation is due to human
cancers and hereditary effects, whereas environmental impacts are negligible.
Emissions of conventional pollutants by the nuclear fuel chain are negligible, if
one allocates energy use upstream and downstream of the power plant to nuclear
(as appropriate for an assessment of nuclear as base load electricity source for
the future).

54
Figure 2. Comparison of damage costs, for fuel chains in the EU,
with kg of Table 2 and emissions of Table 3

m /kWh
0 20 40 60 80 100 120 140

Coal, current
Coal, new

Oil, current
Oil, new
Gas, current
Gas, new
Nuclear, <100 yr
Nuclear, >100 yr

Biomass, high
Biomass, low

PV high PM10,SO2,NOx
PV low CO2
radiation
Wind high
Wind low

Note: “High” and “low” for renewables indicate typical range of estimates of ExternE
[4]. For nuclear only a single technology is shown (French, with reprocessing), but costs
are separated into near (before 100 yr) and far (after 100 yr) future. Note that production
cost of base load electricity in EU and USA is in the range of 25 to 50 m N:K

All this assumes, of course, a mature and stable political system, with strict
verification of compliance with all regulations. Low external costs do not suffice
to allay concerns about accidents, long lived radioactive waste, the right to impose
impacts on future generations, and risks from terrorists and rogue governments;
these issues involve acceptability and defy quantification in terms of external
costs. The following simple example can illustrate why external costs are not the
only decision criterion. Suppose someone invents an energy system that can
13
supply the world’s electricity (roughly 10 kWh/yr) at the bargain price of
$0.075/kWh (about one tenth of typical current prices) – with one little catch:
there is a probability of an accident occurring once every 100 years that will kill
25 000 (roughly the total expected toll for Chernobyl if one multiplies the
UNSCEAR estimate of committed dose by the standard linear dose-response
function).

55
Even at $3 million/life and zero discount rate, the levelized cost of such an
accident has an expectation value of only
2.5 × 10 × 3 × 10 $/(100 yr × 10 kWh/yr) = 0.75 × 10 $/kWh
4 6 13 -4

a mere 1% of this low electricity price. But would people accept such a deal?

Impacts of solid hazardous wastes are impossible to predict to the extent that
they depend on future waste management decisions; the best one can do is
evaluate scenarios. In principle such impacts can be kept negligible by storing
wastes in well-managed leak proof facilities. But will the integrity of the
containers and liners be maintained forever? In case of a leak the most likely
occurrence is leaching into the ground water, and the impacts tend to be limited to
the local zone and could be stopped or corrected, if appropriate measures are
taken. Technologies for alternative methods of solid waste disposal are evolving;
for example, coal ash is increasingly used as additive in building materials. Fly ash
can be stabilized in concrete or glass. For coal none of the externality studies have
succeeded in quantifying physical risks from solid wastes.

For nuclear power, the volume of waste is very small compared to that from
coal but the toxicity is high. Until recently the predominant view was that
nuclear wastes should be deposited in deep geological sites, to be sealed
permanently. Based on certain assumed scenarios for accidental intrusion or
containment failure, similar to other studies, ExternE [3] estimated the
contribution of nuclear waste storage to be about 1% of the total damage cost.

However, in view of all the doubts that have arisen about permanently
sealed geological sites, it may be wiser to envisage storage in depositories that
are guarded and retrievable for the indefinite future (note that the cost, even over
an infinite time horizon, is not excessive, even with low intergenerational
discount rates [10]). After all, whatever the uncertainties for the future, in the
long run technological progress is the most likely scenario, and future
generations will have better means of dealing with radioactive waste, provided it
is retrievable. Is the idea of leaving wastes for future generations really so
shocking, if the alternative is to leave other wastes that are likely to impose even
higher costs? The greenhouse gases from the use of fossil fuels pose a threat the
severity of which is increasingly recognized. And as for long lived solid wastes,
one should not forget that coal ash contains appreciable quantities of toxic
metals, with half life much longer than radioactive waste, not to mention the
natural radioactivity of coal. In view of this situation we believe that the existing
assessments of waste disposal are not very satisfying and that a new set of
comparative studies is needed to put the options for nuclear and for fossil fuels
in perspective.

56
6. Risk comparisons

Since most of the health risk from nuclear power is imposed on people living
in the far future, a simple number for the global YOLL is not very instructive. For
another perspective let us look at the implications of using nuclear power on a
large scale. The population of the world is about 6 billion and the electricity
consumption 12 000 TWh/yr. Both will increase, especially the latter, although
saturation can be expected eventually. Technologies will certainly evolve and
nuclear fission reactors are unlikely to be more than a stopgap, perhaps for a
century or so, until cleaner sources of energy mature. Since details do not matter
for the following argument, we take simple round numbers.

Let us suppose a simple “100 year scenario” where the above nuclear fuel cycle is
used for 100 years to produce 2 × 10 TWh/yr, for a world population of 10 . The
4 10

total public dose for the French nuclear fuel chain is 12.5 person·Sv/TWh; most of
this dose is imposed globally rather than being limited to a local zone. This dose
would imply a dose rate of
Total dose rate = 12.5 person·Sv/TWh × 2 × 10 TWh/yr/10 persons = 25 Sv/yr
4 10

if the entire dose were incurred immediately. However, only ten percent of the
dose from the production of electricity is incurred during the first 100 years. The
precise time distribution of the total dose rate from the “100 year scenario”
would be difficult to calculate, because of the large number of different half
lives. But again a rough order of magnitude estimate is sufficient, and we simply
take 10% of the total dose rate to be imposed on the population living during this
“100 year scenario”
Total dose rate first 100 years = 2.5 Sv/yr.

This can be compared to background radiation from other sources that


people are exposed to. A typical value is in the range of 2 to 3 mSv/yr, including
medical x-rays and an average value for radon in buildings. Another interesting
comparison is with the dose rate due to cosmic radiation at sea level, 260 Sv/yr,
because this is the very minimum all of us are exposed to and protection from
which is not practical. Thus the “100 year scenario” would increase the
background exposure by, very roughly, 1% of the cosmic ray background at sea
level. Generations living beyond those 100 years would of course also be
exposed, but at a lower rate.

In Figure 3 we compare on a logarithmic scale the YOLL due to fatal


cancers from the 2.5 Sv/yr dose rate of this “100 year scenario” with several
other risks of everyday life (to get numbers that are easier to comprehend, we
convert to days of life lost). Also included is the mortality risk due to another
potentially significant source of renewable energy, namely waste incineration,

57
for a scenario where all the municipal solid waste (500 kg/yr per person) is
incinerated with emissions equal to the latest Directive of the EU, effective after
2001. The risks from these two controversial technologies, waste incineration
and nuclear power, are very small.

Figure 3. Comparison of some public health risks

Days of life lost, per person per lifetime


0.01 0.1 1 10 100 1 000

Car accidents, France

Pedestrians killed by cars, France

Lightning, France

Fatal hunting accidents, France

Deaths due to terrorism, France

Air pollution in Paris, if increase by 10%

Waste incineration

Radiation background 2.5mSv/yr

Nuclear power for world for 100 yr

Note: Scenario for waste assumes all municipal solid waste is incinerated, with emissions
equal to the latest EU regulation. Scenario for nuclear assumes the entire current world
consumption of electricity is supplied for 100 years by the current technology of France.
Data for the first 5 items are from Frémy & Frémy. [5], for air pollution in Paris from
Rabl [14] and for waste incineration from Rabl et al. [12]. Error bars express uncertainty as
estimated by Rabl & Spadaro [13].

7. Conclusion

Based on the results of the ExternE Project we have compared the public
health risks and the damage costs from the routine operation of energy systems.
Generally the loss of life expectancy per kWh of electricity is much higher for
fossil fuels than for modern nuclear plants and for most renewables, and among
fossil fuels it is much higher for coal than for gas. However, for each energy
source there is much variation between different installations of a given energy
source; for example, the emission of harmful pollutants from coal power plants

58
built after January 2000 will be an order of magnitude lower than typical
emissions from existing plants in the USA and EU during the nineties.

The health effects of PV (photovoltaics) and wind arise from the production
of the materials rather than during operation of the plant. Among the renewables
only biomass combustion entails relatively large health risks, due to the emission
of PM10, SO2 and NOx. The health risks from the routine operation of the nuclear
fuel chain are very small, even though they have been summed over a time
horizon of 100 000 years. Since the meaning of risks imposed over thousands of
years is difficult to comprehend, we have illustrated them by considering a
scenario where the entire electricity of the world for 100 years is produced by
nuclear: the incremental collective dose is only one percent of the cosmic
radiation background.

Even though the uncertainties are large (the numbers could be about 4 times
smaller or larger), they have relatively little effect on the comparison between
energy systems, except for nuclear, because they affect the impacts of the air
pollutants in more or less the same way.

Acknowledgments

This work has been supported in part by the ExternE Project of the
European Commission DG Research.

REFERENCES

[1] R.W. Brode, J. Wang, User’s Guide for the Industrial Source Complex
(ISC2) Dispersion Model, Vol. 1-3, EPA 450/4-92-008a, EPA 450/4-92-
008b, and EPA 450/4-92-008c. US Environmental Protection Agency,
Research Triangle Park, NC 27711, 1992.

[2] R.G. Derwent, K. Nodop, Long-range Transport and Deposition of Acidic


Nitrogen Species in North-west Europe, Nature 324, 356-358, 1986.

[3] ExternE 1995. ExternE: Externalities of Energy. ISBN 92-827-5210-0.


Vol. 1: Summary (EUR 16520); Vol. 2: Methodology (EUR 16521);
Vol. 3: Coal and Lignite (EUR 16522); Vol. 4: Oil and Gas (EUR 16523);

59
Vol. 5: Nuclear (EUR 16524); Vol. 6: Wind and Hydro Fuel Cycles (EUR
16525). Published by European Commission, Directorate-General XII,
Science Research and Development. Office for Official Publications of
the European Communities, L-2920 Luxembourg.

[4] ExternE 1998. ExternE: Externalities of Energy. Vol. 7: Methodology 1998


Update (EUR 19083); Vol. 8: Global Warming (EUR 18836); Vol. 9: Fuel
Cycles for Emerging and End-use Technologies, Transport and Waste
(EUR 18887); Vol. 10: National Implementation (EUR 18528). Published
by European Commission, Directorate-General XII, Science Research and
Development. Office for Official Publications of the European
Communities, L-2920 Luxembourg. Results are also available at
http://ExternE.jrc.es/publica.html. For a summary in French, see A. Rabl,
J.V. Spadaro, Les coûts environnementaux de l'énergie,. Rapport ADEME,
action EP3, et Commission Européenne DG12, Contrat JOS3-CT95-0010.
ARMINES/École des Mines de Paris, Paris, 1997

[5] D. Frémy, M. Frémy, Quid 1997, Éditions Robert Laffont, Paris, 1996.

[6] IPCC 1995, Climate Change 1995: Economic and Social Dimensions of
Climate Change. J.P. Bruce, H. Lee, E.F. Haites (Eds.), Contribution of
Working Group III to the Second Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge University Press,
Cambridge.

[7] W. Krewitt, F. Hurley, A. Trukenmüller, R. Friedrich, Health Risks of


Energy Systems, Risk Analysis, Vol. 18(4), pp. 377-384, 1998.

[8] ORNL/RFF. External Costs and Benefits of Fuel Cycles. Prepared by Oak
Ridge National Laboratory and Resources for the Future. Edited by Russell
Lee, Oak Ridge National Laboratory, Oak Ridge, TN 37831, 1994.

[9] C.A. Pope, M.J. Thun, M.M. Namboodri, D.W. Dockery, J.S. Evans,
F.E. Speizer, C.W. Heath, Particulate Air Pollution as a Predictor of
Mortality in a Prospective Study of US Adults, Amer. J. of Resp. Critical
Care Med., Vol. 151, 669-674, 1995.

[10] A. Rabl, Discounting of Long-term Costs: What Would Future Generations


Prefer Us To Do? Ecological Economics, Vol. 17, p. 137-145, 1996.

[11] A. Rabl, N. Eyre, An Estimate of Regional and Global Ozone Damage


from Precursor NOx and VOC Emissions, Environment International,
Vol. 24, 835-850, 1998.

60
[12] A. Rabl, J.V. Spadaro, P.D. McGavran, Health Risks of Air Pollution from
Incinerators: a Perspective, Waste Management & Research, Vol. 16,
pp. 365-388, French version published as “Effets sur la santé de la pollution
atmosphérique due aux incinérateurs : une perspective”. Déchets, Sciences
& Technique. No. 9, pp. 4-22 (1998).

[13] A. Rabl, J.V. Spadaro, Environmental Damages and Costs: an Analysis of


Uncertainties, Environment International, Vol. 25, 29-46, 1999.

[14] A. Rabl, Les bénéfices monétaires d’une amélioration de la qualité de


l’air en Ile-de-France, Pollution Atmosphérique, janvier-mars 1999,
pp. 83-94, 1999.

[15] R.D. Rowe, C.M. Lang, L.G. Chestnut, D. Latimer, D. Rae, S.M. Bernow,
D. White, The New York Electricity Externality Study, Oceana Publications,
Dobbs Ferry, New York, 1995.

[16] D. Simpson, Photochemical Model Calculations Over Europe for Two


Extended Summer Periods: 1985 and 1989. Model Results and Comparison
with Observations, Atmospheric Environment, Vol. 27A, p. 921-943, 1993.

[17] J.V. Spadaro, Quantifying the Effects of Airborne Pollution: Impact


Models, Sensitivity Analyses and Applications, Doctoral thesis, École des
Mines, Paris (France), 1999.

[18] J.V. Spadaro, A. Rabl, Estimates of Real Damage From Air Pollution:
Site Dependence and Simple Impact Indices for LCA, International J. of
Life Cycle Assessment. Vol. 4 (4), 229-243, 1999.

61
Session 2

EXTERNAL COSTS OF ENERGY/


ELECTRICITY LIFE CYCLES
(RESULTS FROM RECENT AUTHORITATIVE STUDIES,
LESSONS LEARNED, UNCERTAINTIES, GAPS)

63
A LIFE CYCLE PERSPECTIVE OF COAL USE

Louis Wibberley
BHP Sustainable Technology, Wallsend (Australia)

Context

• Background
– negativity towards coal has been based on superficial “burner tip” comparisons -
a misleading basis for comparison
– policy dilemma - how to meet the development needs of the world in a
sustainable, affordable manner
– coal is expected to underpin future energy demand in many parts of the World (large
reserves, diversity of supply, stability of price, ease of storage)
– reduced proportion of the total energy is expected, but the overall consumption will increase
– renewables need a base load energy source
• Australia has 2 related initiatives for understanding and improving the
role of coal in the transition to a more sustainable society
– a life cycle analysis of steel and electricity production for a range of energy
sources and technologies (Australian Coal Association, WCI and others)
– Cooperative Research Centre for Coal in Sustainable Development (CCSD)
– a 7 year program involving industry, Government and research organisations
– fundamentals, technology and holistic analyses of most aspects of coal use
Minerals - Newcastle Technology Centre

65
Approach - LCA component

• To carry out streamlined life cycle analyses for a wide


range of steelmaking and electricity process chains
– cradle-to-gate inventory analysis
– limited to resource energy, GGE, NOX, SOX, particulates, solid waste and water
– using (initially) Australian data
– focus on energy/technology comparisons, improvement opportunities,
detailed understanding of the production chain
• To use streamlined impact assessment
– aggregation limited to GGE (taken as the main proxy for impact)
– restricted to environmental burdens, no economics assigned
• LCA chosen to help focus on assessment of
improvement options
• Broader framework for assessment of socio-economics
is being considered in the CRC program
Minerals - Newcastle Technology Centre

ExternE values - coal based electricity in Europe


(mECU/kWh)
YOLL Global
Warming All
Country Total Comments
GHG Mid Other
SOx N Ox Other Total 3%

U. K Deep mine PF ESP,


U.K. 6.1 10.5 2.9 19.5 15 7.5 42
FGD Low Nox No SCR

2.9
Ge rmany 6. 3 1.2 10.4 14.3 5 29. 7 DENOX FGD

SCR for NOx > 90%


reduction FGD for SOx >
Swe de n 0. 079 0. 366 0. 036 0. 481 13.2 4.419 18. 1 88% reduction Electric
filter for PM > 99.9%
reduction Co generation

YOLL = Years of life lost converted to economic terms Reference: EUR 18528 – ExternE- Externalities
* Other (= includes morbidity costs of TSP, SOx & NOx, and accidents ( accidents minor contributor) of Energy Vol. 10 National Implementation

All Other = cost of impacting crops, ecosystems, materials, noise, aquatic systems & aesthetics
*Mid 3% GHG: A discount rate is applied to future impacts of global warming events

Minerals - Newcastle Technology Centre

66
1

Historical perspective - iron & steelmaking


Impressive process improvements have been made by the steel industry over time,
by both breakthrough and incremental technology development

Low bloomery (charcoal)


1000 Liquid iron (charcoal BF)
CO2 t/t steel bar

Wet puddling
100 Coke BF
Hot blast BF
Bessemer steelmaking
10 Open hearth
BOS
Continuous casting
1
Recycling
Integration
? New technology
0
0 500 1000 1500 2000 Renewables
Year

67
Steel GGE (t CO2-e/t cast steel)

• Why is a systems or holistic approach


so important?

Minerals - Newcastle Technology Centre

LCA system

Emissions
to air
Slag
(cement credit)

All processes
Resources
in ground involved in the Functional unit:
production of cast 1 t steel
steel
Offgas
(electricity credit)
Emissions Emissions
to land to water

Minerals - Newcastle Technology Centre

68
Displacement credits - slags
a) BF slag processing system BF slag cement
(basis 3,500 kg hot metal) GGE 60 kg CO 2-e
(equivalent to 1,000kg
of Portland cement)
No technical or
60 kg CO2
economic issues

1,000 kg Often limited by


attitudes
Blast furnace Slag grinding
A product
b) Cement system Portland cement stewardship
GGE 1,020 kg CO2-e
1,020 kg CO 2 issue for both
coal and steel?
Limestone and 1,000 kg
shale quarrying

Cement plant
(includes clinker grinding)

Minerals - Newcastle Technology Centre

Steel GGE (t CO2-e/t cast steel)

0.0 1.0 2.0 3.0 4.0 5.0

BF - BOS Existing

Corex - BOS

Midrex - EAF Gas based DRI

New technology Emerging coal technology

Net GGE Electricity credit Slag credit


Minerals - Newcastle Technology Centre

69
Blast furnace is only one source of GGE
coal supply
coke ovens
sinter plant
hot blast
blast furnace
power plant
BOS
electricity
aluminium
transport
other
byproducts
gross GGE
slag credit
electricity credit
net GGE
0.0 0.5 1.0 1.5 2.0 2.5 3.0
GGE (t CO2-e/t cast steel)
Minerals - Newcastle Technology Centre

Improvement opportunities

100
Reduction in GGE (%)

80

60

40

20
p

n
its

y
l

ra

io
ta

as

og
M
ed

at
Sc
en

CB

om

ol

gr
Cr
m

hn
Bi

te
e

ag

c
cr

In
te
Sl
In

w
Ne

Minerals - Newcastle Technology Centre

70
Coal bed methane (CBM)
Turkey
• World total 30Mtpa?
Russia
• only 5% utilisation?
USA • ~50% as MVA for underground
China mines
Australia • biggest GGE benefit from
India oxidation, power gives small
Indonesia additional benefit
0 5 10 15 20 25
Methane content (m3/t)

Pre-drainage methane
Underground (35 - 90% CH4)
coal mine CBM MVA
(0.2 - 0.8 % CH4)
Ventilation air

Minerals - Newcastle Technology Centre

CBM utilisation at Appin and Tower

• 94MWe using 1MWe


gas engines

• 160kt/a CH4 utilised


(pre-drainage gas, some
MVA used for combustion)

• 3Mt CO2-e avoided


annually

Minerals - Newcastle Technology Centre

71
Charcoal - limited applicability …
Charcoal trials at Corrimal

Cost $350-500/t
Niche markets already economic
(e.g. recarburiser is 10-20kt/a in Australia)
Biomass to generate electricity is a more
effective approach
♦ less transport
♦ more flexibility in biomass type
Minerals - Newcastle Technology Centre

Minerals - Newcastle Technology Centre

72
Historical perspective – power generation
g
in
at
y r oc s
da s cip ine ine
ra ate e
Fa ner icity Re ng rb in
ge ectr 882 m e m tu tu
rb
1 tea ea kW
9000 el s St e d am ,000
te
884 duc fs 50 t,
8000 1 tro o - ea ng
se 10 rh iri
GGE (kg/MWh)

in lu le pe on f
a a u
7000 rs sc , s nsi
ive ing ale pe
6000 Un eas sc sus ifi
ed
cr d ,
In
e
as ces , un
e y
5000 r
n c rn
a cit ies n,
5 I fu pa og
-3 all ca nol ratio les
4000 2 0 w r h g ab
19 ter ge ec te
ar l t in ew
3000 wa 0 sLs c oa SC, ren
7 n n U ic
19 sig ea C, ist
2000 de Cl GC erg
I n
sy
1000
0
1800 1850 1900 1950 2000 2050
Increasing scale
Superheat & pressure

Electricity GGE (t CO2-e/MWh)


0.0 0.2 0.4 0.6 0.8 1.0
Conventional Coal

NG C-C Range
Gas
LNG C-C Range

IGCC
Clean coal
Future coal

Hydro Range

Photovoltaic Renewables
Wind, biomass

Nuclear ?
Net GGE Ash Credit
Minerals - Newcastle Technology Centre

73
Improvement opportunities

100
Reduction in GGE (%)

Basis: 36% NTE


80

60

40

20

ng
e

d
r-t ass

Em ical
pe ical
ya tal

al
M
us

ne
gi
rm
CB
en

om

rit

bi
it

er
sh

he

tra ercr
m

m
rc
Bi
re

Co
Fl

p
c

la

su
Su
In

So

Ul

Minerals - Newcastle Technology Centre

Reduction options

Option Change in GGE reduction


efficiency* (%)
Incremental improvements 36 38% 5
Replacement
Old coal with new 26 40% 25
Supercritical pf 36 40% 10
Ultrasupercritical pf (now) 36 42% 15
Ultrasupercritical pf (future) 36 50% 30
Emerging IGCC etc 36 50% 30
Flyash to cement 5-7
Biomass-coal 5-15
Solar-coal 10
* gross, sent-out
Minerals - Newcastle Technology Centre

74
Synergies with renewables

Biomass co-firing Solar thermal


35% biomass conversion 30-40% solar conversion
efficiency (20% for dedicated) efficiency (13% for PV)
Coal can promote uptake and efficient utilisation of renewables
Coupling of renewables and fossil energy research is essential

Minerals - Newcastle Technology Centre

Biomass-coal generation

Partnership with a Future:


Coal and Bagasse
Dual-Fired Belle Vue Power Plant
in Mauritius
• Guadeloupe, Reunion and Mauritius have installed 6 x 70 MWe
dual fuel power stations:
– bagasse (6 month season)
– coal (when bagasse unavailable)
• Provide electricity throughout year, while maximising use of
renewable energy (biomass)
• economic and social benefits
• enable more efficient plants to be built
Source: Good News from Coal, WCI, Nov 1999
Minerals - Newcastle Technology Centre

75
Solar-coal generation?

• Several technologies
have been proposed
– 130 MWe per km2
• Lowest cost routes to
solar electricity
– A$80/MWh @ 100 MWe
• Demonstration plant of
3 MWe (av) under
consideration

Minerals - Newcastle Technology Centre

Power generation – water use


10
Cooling Water (t/MWh sent out)

0
1880 1900 1920 1940 1960 1980 2000 2020 2040

Increasing turbine efficiency


Minerals - Newcastle Technology Centre

76
Power generation - water use

• Water consumption for power generation depends


upon the cooling technology used and the efficiency
of the conversion of steam to electricity in the turbine
• Majuba power station in South Africa
Water Efficiency
consumption (%)
(m3/MWh)

Units 1-3 (dry cooling) 0.2-0.4 ~33

Units 4-6 (wet cooling) 2.0-2.5 ~37

Source: African Energy Vol. 1, No.3, 1999


Minerals - Newcastle Technology Centre

Final remarks
• Streamlined LCA has provided improved data and
understanding of issues for coal and alternatives
– for GGE, the difference between coal and alternatives has been shown to
be less than previously reported
– still many information gaps
• Many opportunities for improvement throughout the coal
chain - from coal in the ground, through to waste disposal
– slags and flyash are not used to maximum benefit
– oxidation/utilisation of CBM
– substantial improvements are available through incremental changes to
“conventional pf” technologies (in addition to the new technologies)
• Coal can underpin the introduction of renewables
– higher overall efficiency (and more cost effective than stand alone
options)
• Further studies are underway
– combining LCA, technology evaluations, socio-economics
Minerals - Newcastle Technology Centre

77
WELL-TO-WHEEL ENERGY ANALYSIS STUDY

Jean Cadu
Shell, London (United-Kingdom)

(http://www.transportation.anl.gov)

What is a Well-to-Wheel Analysis?


• Systems Approach
• Assessment of energy consumption and greenhouse
gas emissions

Well-to-Tank

Tank-to-Wheel

79
80
81
82
83
84
85
86
87
88
89
FROM LIFE CYCLE ANALYSIS APPROACH TO
MONETARISATION OF THE IMPACTS:
AN EVALUATION IN TERM OF DECISION PROCESS1

Marc Darras
Gaz de France, Paris (France)

The decision process both of policy-makers and firms is never limited to a


strict cost-benefit analysis. In the field of energy, a whole chain of decision is
needed to allow a good and efficient use of any energy form, from the long-term
investment of exploration at the upper-stream, to the development of process and
the affordability analysis by the final consumer. Every decision needed along this
chain affect the supply, and the security of supply, the quality of energy, the
economical and social development both directly by the activity of the energy
sector, but mostly indirectly through the benefit of energy, and finally the
environment in a very large manner. These decisions are taken having a view of
all these non-economic effects, even if the framework of decision is not always
2
clearly organised to provide the highest level of rationality.

Following Rio Conference, a strong emphasis has been put on sustainable


development which promote a social, economic and environmental develop for
the present and future generation. The programme Agenda 21 proposes goals in
further multiple direction than the ones above-mentioned.

The life cycle analysis offers a new tool in the direction of integration of the
multiple physical flows of matters around industrial processes and products,
including the usage stage. Simultaneously, it stresses the multidimensional
impact of the human activities, first in term of physical flows, but then in term of
impact on the goals stated above.

In this perspective, integrating framework for decision are needed,


especially at the more global level. One of the proposed frameworks is the

1. This article presents the views of the author and is not a position either of Gaz de
France or the International Gaz Union.
2. Concept still to be defined.

91
monetarisation and cost-benefit analysis, as an extension of the quantitative
economy. The multi-dimensional choice, unfit for strict order choice, is reduce
to one variable.In the presentation below, because of the importance of this
approach in connection with Life Cycle Analysis and public policies, we
concentrate on the analysis of this framework. We will show that it might shed
light on the decision process, especially if associated with its uncertainties, its
limits in term of process or criteria taken into account, the indication of other
social criteria not taken into account, and other method of evaluation. At the end,
this is mainly a source for a better multi-criteria decision process.

1. What is it all about?

Industrial activity is by nature a transformation activity involving materials


from various resources; goods or services are produced. During this process
wastes are generated in a gaseous, liquid or solid form.

The traditional operational field of a firm is economy, where goods are


exchanged on the basis of values fixed within a market. The increased use of
limited natural resources and the awareness of the impact of industrial or domestic
waste and effluents have revealed another field, which is the environment.
Consequences outside the economy must therefore be integrated to the human
activity assessment: the economists have named them externalities.

For the Firm, these externalities often exist in two ways: external to its
economy field, external to the activity field of the Firm. It really means taking into
account on one hand the diminution of resources (raw materials, space, etc…) and,
on the other hand, the impact to which the process effluents contribute.

The Firm must then consider theses environmental externalities in


alternative positions:

• In the face of the evolution of a society wishing to take into account


environmental issues when setting its options; for the Firm, this is part of
the strategic positioning of its activity, of its products.

• In the face of its own production and investment options, either with the
point of view of exemplarity – which is a strategic option, or plainly
within the frame of regulatory and social constraints.

In either term of the alternative, rationality must be preserved in order to


optimise industrial production in a way adapted to clarified social options.

92
The long practice of cost/benefit analysis in firms, as well as a search for
global optimal economic balance have led economists to try and approach
environmental impacts as if it were a cost: it is the valuation of impacts or
monetarisation. This approach has two anticipated advantages:

• The attenuation of impacts of different natures, incommensurable (how


can one say that a noise nuisance is of less importance than a spoilt
landscape) within a common measurement unit, which can further be
merged to economic costs in the process of optimised calculations.

• The assessment of the cost of the environmental impact of an activity,


which would enable a polluter-payer approach.

2. What are the methods?

Two environmental cost assessment methods are commonly identified: the


two-tier methods, and the direct methods.

The two-tier methods, or dose-response analysis, are the results of an


analytic approach:
Emission Dispersion Impact Evaluation

The direct methods progress directly through an identification of the


economic value of an impact, neglecting its causal chain. Several can be
identified:

• Cost of damage repairs (medical care, double glazing, …).

• The transport costs method, in which the value of an environmental


“free” asset (park, forest) is assessed to the value of the transport
expenses required to benefit from it.

• The hedonist pricing method, in which after the survey of various data
related to an asset on different markets, the shares allocated to each
environmental factor (noise, landscape, … impact on the price of
housing) are valued.

• The contingent pricing method, which processes directly through


questionnaires to asses the price each agent attaches to an environmental
asset (willingness to pay, to receive).

93
Each one of these methods presents several alternatives which have been
worked around since the early 70s.

In reality, and here for the purpose of assessment of the valuation stage, the
merits of the direct or two-tier method are not significant. Indeed, the crucial
features of these approaches are:

• The identification of any impact which will require evaluation (noise


pollution, morbidity, odours, …).

• The content of the evaluation indicator, whether a valuation refers to a


measuring unit or is globalised.

In the dose-response method, physical pollution indicators (waste volume


and substance, particle concentration, …) are initially available, which are
converted into impact through a dose-response function (toxicity, eco-toxicity),
then through a monetary valuation of the response. The second approach, the
direct method, also allows to deal with impacts which cannot be quantified
through an intensity scale (aesthetic pleasure, …).

The effective content of the value indicator (= value) of an impact is a


crucial process. Indeed, this value can represent, capture, according to the
circumstances:

• The value of the assets within the economic scope directly supported by an
industrial plant (cost to reforest quarries, now internalised by regulations) or
collectively (cost of façade restoration as a pollution indicator).

• The value of impacts which slip out from the economy scope because they
are not connected generally to markets: “option value” which referred to
the preference for a delay of decision, or “quasi-option” which enables to
maintain reversible options, “existence value”, “legacy value”.

In the latter category, individual or collective choices have made their


appearance, tied in with sociological impacts (free access to the forest, sacred
aspect of worship places, protection of whales, …), but as well impact in nature
related to the precaution principle (long term storage of radioactive waste,
biodiversity, …) very far apart from the exchange of goods. These impacts are
essential in the decision process, or more precisely in the process of taking
possible decisions with all the stakeholders, even if the rationality is to be
clarified in some cases. The valuation indicators do not then necessarily capture
all of these concepts.

94
3. Environmental impact valorisation difficulties and limits

The experience gained on the environmental impact valuation methods over


more than twenty years, as well as the methodological consideration that
surrounds this approach, allows delimiting a number of difficulties, which
should be kept in mind when assessing the outcome of such surveys. We have
compiled these difficulties under a few headlines, without any specific order or
relative significance for each of the points.

3.1 Domain of assessment

The externalities monetary valuation system leads to questions about the


nature itself of what the valuation actually represents. Therefore, it appears that
there are different levels according to the valuation mode.

Firstly, and when the values are obtained either from market costs (façade
restoration, structural maintenance, …), or from actual expenses (travel costs),
valuations are commensurable with actual market economic costs.

However, when they are obtained through contingent cost methods, what
consistency do these bear with economic costs? Let us go back to the previous
example: the questioning about the willingness to pay to have clean façades does
not necessarily lead on to the value of the restoration. If actors are well
informed, involved in such type of operations, their answer will be biased by
their knowledge of the actual economic cost, and they will tend to come close to
it. However, an actor away from this problem will determine a cost in relation
with the importance such an operation has for him, hence linked to his income,
and, on another hand, to his sensitivity to the aesthetic of façades and other
competing needs/wishes: it will then be a matter of psychological sensitivity.
The connection with an economic cost becomes diluted, and the combination
more precarious.

The experience acquired on contingent valuations (willingness to pay, to


receive) shows the great variety of appraisal and the sensitivity to psychological
factors: structure of the questionnaire, type of questions, … Finally, there are
impacts which escape valuation since they remain inconceivable to the individuals
3
questioned: can forest cease to be a free asset when one lives in Finland?

Therefore, there is a necessity to clearly make a distinction between


valuations actually emerging from the economy sector from those conveying or
capturing in part a social or psychological value (opportunity of a CMU unit,

3. Not being a Finn, it is just a guess on my own.

95
Contingent Monetary Unit, which will not mix will true monetary exchange or
only on certain evaluation phases?).

3.2 Aggregation of the effects

For a method, within an analysis:


emission transfer impact valuation,
the valuation stage will allow to aggregate the impacts. This is a delicate procedure.
Indeed, some joint emissions have an impact that is different from the sum of
impacts of each emission, either because the impact has been added twice, or
because the joint effect is of a different nature. In this way, a recent survey of urban
pollution was found difficult to wrap up Thus must be monitored the consistency
between types of impacts identified, the estimate of the dose-response function and
of the impacts assessed. Finally, the same emission can partake in several impacts of
strongly diversified natures (particles produced by motor vehicles can lead to
respiratory problems and to the soiling of façades). The valuation is no more
connected to the emissions, but to a complex application based on the emissions.

On another hand, the mode of identification of values, and particularly using


the hedonist methods, leads to limit the use of these values in domains of impact
of the same order of magnitude than with the observations that have been used as a
basis for the identification. Any extrapolation would be hazardous since there is no
identification of causalities. This is actually the case for any empirical rule
statistically determined and based on a given set of observations.

3.3 Temporal and spatial aggregation

The traditional economic calculation uses temporal aggregation using


present value ratio, which shows the preference for the present. There is massive
literature dealing with the effective value to be considered. However, this
approach requires that only actions be considered that do not include any
“irreversible” effects, since present and future actions must be interchangeable
by nature.

Moreover, when considering environmental effects that might be delayed or


spread over a very long period (greenhouse effect, long-life radioactivity, …) the
present value estimate of any effect becomes negligible at a 50-100 year time-
horizon. The issue here is no more taking a risk as in a traditional cost-benefit
present value calculation, but more an ethical problem towards the future
generations. Will this point be lifted just counting on a technical evolution to
come, thus allowing future generations to efficiently compensate technically and

96
economically for an impact passed on by the present generation? Finally for
certain distinct class of risks, impacts, of major importance even if not with a
quantified probability of occurrence, a precautionary principle forbids this type
of temporal aggregation.

Inter-temporal aggregation sets a second problem due to an ambiguity. We


have indeed determined that contingent valuation is more a social appraisal
indicator than a value in itself. It therefore rests on the social distribution of the
values, and on the connected rights and duties. This distribution of the values
varies in time, and even more when we move away from fundamental ethical
values towards preferences, tastes … It is therefore impossible to trim down
these future preferences to the present ones, hence to assess them. In this
context, an option based upon identified present social values and leaving major
options for the future open is the only alternative.

Spatial aggregation raises a problem of similar nature, particularly when


there are spatial heterogeneities. As an example let us take the impact
assessment of the establishment of a mine. If it appears that this establishment
must take place in a low-income area, then the impact based on the cost of
housing is assessed to be low. A similar problem exists when the morbidity
based impact is valued through the economic cost of an economic agent
(education, production to be, …): the impact is then smaller in less developed
areas. It is therefore necessary to check that the environmental impact design to
help the decision process does not introduce an heterogeneity between the
populations, eventually correct in economical terms, but inconsistent for the
social choice purpose.

3.4 Accuracy of the assessment

The accuracy of the global assessment rests on numerous aspects:

• Variation/accuracy of individual impact assessments.

• Accuracy of the assessment of the emission, transfer and response (dose-


response function).

• Good sampling of all potential impacts, particularly when the analysis is


based on emissions.

• Appropriate identification of potential cumulative effects.

97
Each of the above stages is subject to a more or less obvious ambiguity.
This uncertainty must hence be estimated, taking into account the cumulative
effect through the valuation procedure.

4. What is the use of monetary evaluation of externalities for the firm?

As previously stated, the externalities monetary assessment appears within


two situations:

• A global context in terms of social valuation in which is encompassed


the activity of the firm; this is the case for the full cost of energy.

• A micro-economic context, directly linked to the activity of the firm.

Taking into account the actual context of externalities monetary valuation, we


will now study how this approach can be better used within the firm decision process.

4.1 Global appraisal of the environmental impact of energy

Global appraisals of the environmental impact of energy have been


attempted over the recent years by several organisations, and notably by the
European Union on the cost of electricity (ExternE).

The approach of externalities seems to suggest here a reasonable framework


for the debate. What about it? On the basis of the above analysis, is appears that:

• The traditional economic approach calculates an economic optimum for


the exchange of goods; taking into account full externalities should lead
to an economic under-optimum, or to an optimum under “social”
constraints.

• The procedure is kept within the bounds of economy as long as it


integrates only the share of environmental effects (external or internal)
which is relevant to the economy, otherwise an added monetary flux is
added in the system (contingent monetary unit flux).

• When monetarisation is extended to effects beyond this economic scope,


the whole of the environmental effects is still not taken into account: it is
therefore a light shed on only a part of the environmental impact.

• The appraisal of environmental impact hence requires further


investigations complementary to monetary appraisals.

98
• Concerning the monetary valuation, the value uncovered must be
associated with a survey covering the uncertainty and sensitivity to the
monetary unit values selected.

In fact, we are finally confronted to a bundle of indicators, some of which


are given in monetary or quasi-monetary terms. In order to analyse the
positioning of a decision, each one of these indicators must be assessed so that
the most relevant opinions can be expressed and help improve the consistency of
the decision process. One example would be to take into account the outcome of
a strict economic survey (including, of course, the environmental effects which
have been internalised according to the regulation or voluntarily by the firms, the
consumers, ...), then include one or another environmental effect depending on
what the stakeholders can valorise, may negotiate or might take into account.
The whole set will act as a support to the decision process.

However, reasoning over the consistency of a decision options can be


compared with the above remark to a multi-criteria analysis, in which irreducible
data are compiled and clearly weighed in accordance with the options set by the
various parties (firms, administration, consumers …) through the procedure
aiming at defining the weight for each parameter. One will notice that this
procedure can be implemented directly at impact identification level, associated
or not to an extensive valuation. Furthermore, it forces the reflection on the basis
of the weights of each impact.

Finally, let us note that there is no rational link between a monetary


evaluation of externalities which lay outside the economic domain, and the value
of a tax which is directly affected in the economy stricto sensu.

4.2 From life cycle analysis to monetary valuation of a specific activity of a firm

The review of a specific activity of the firm in terms of environment can be


the basis of a life cycle analysis. There again, the remarks made earlier relatives
the consistency of a cost-benefit using monetary valuation of impacts in the
decision process following the life cycle analysis. Therefore it is more
interesting to look at the various option opened by a Life Cycle Analysis.

First of all, it should be pointed out that numerous industrials underline the
gain achieved by the simple analysis of the material flux within the processes, or
for products, a sharper approach than the traditional globalised cost-benefit
assessment.

Insofar as various valuations appear which must be taken into account


simultaneously, a multi-criteria procedure can be used. Practically, it is already a

99
more or less formal approach, because a number of criteria are integrated such as
the maintenance of a process, the development potential, … which are not
always fully economically assessed in terms of advantages. Then, noise, air and
water impact can be added to this approach as independent criteria. It is for
instance one of the basis of eco-efficient approaches promoted by the WBCSD.
Furthermore, such an approach may be a good basis for a multi-stakeholders
decision, within and outside the firm.

This approach leads to study then the processes in terms of cost-efficiency,


and no more in terms of absolute cost-benefit. The environmental externality is
integrated to the constraints of the procedure as a standard to comply with: the
various processes or products can then be compared in terms of cost, the costs
being related to actual economic values. This approach is the cost efficiency
approach. It is to be underline that it is a common industrial practice under the
format of technical clauses (and sometimes administrative) within the supply
specifications for instance. As an add-on to this approach, the sensitivity of the
result to the level of the standard can be tested.

5. Summary of the conclusions

With regards to the consideration of the stakeholders for their options, and
even if the apparent simplicity of:
cost (economic cost + monetarised externalities) – benefit
seems to be a fair option, the valorisation of externalities is tainted with too many
uncertainties for such an approach to be regarded as rational. Alternative
approaches must be considered: multi-criteria methods on the impacts, cost-
efficiency approach, … In all cases however, the stage where only a limited
numbers of environmental effects taken into account must be selected is
unavoidable: this selection is a decision to be made in accordance with the
environmental approach defined by the firm, the priorities of the authorities.

The inter-temporal aggregation leads to uncertainty and may rise ethical


issues which are not solved by the standard economy procedure (present value).
On the long term, the value of the inter-temporal rate of actualisation should be
selected very carefully. Furthermore, when the risk corresponds to a major
change, such as corresponding to the application of the Precaution Principle, the
aggregation is no more founded.

Confronted to an approach based on the monetary evaluation of


externalities, for example the environmental cost of energy, each and every
constituent must be scrutinised, and the robustness of the outcome must be
assessed according to the variability of the elementary valuations. Moreover, and

100
for those impacts, which remain outside the economic scope, it is necessary to
study the benefits of the monetary valuation carried out. The whole operation
must allow the stakeholders to position themselves, to negotiate on a better
understood rational basis.

Negotiations based on the multi-criteria methods can be consistent and


interesting insofar as they can help to bring to light and define the context. Such
approach can take full advantage of the result of a Life Cycle Analysis.
However, the implementation of such methods requires a strong involvement of
the actors concerned to develop the motivation behind the weights given to the
various criteria.

There is no link between the monetary value of externalities of non-


economic nature, and the value of the tax designed to internalise externalities.
Furthermore, the geographical distribution of impact should be taken into
account in the repartition of the corresponding regulation in order to develop the
adequate adaptation. This is particularly important in recycling the revenue, if a
taxation is applied.

BIBLIOGRAPHY

S. Faucheux, J.F. Noël, Économie des ressources naturelles et environnement,


Armand Colin, 1995.

D. Pearce, Les entreprises et les valeurs de l’environnement, in Environnement et


choix économiques d’entreprise, C. Stoffaes, J.M. Richard, ed. Intereditions, 1995.

M. Holland, M. O’Connor, J. O’Neill, Casting Environmental Damage. A Critical


Survey of Current Theory and Practice, and Recommendations for Policy
Implementations, Université de Versailles, St-Quentin en Yvelines, C3ED,
Lancaster University, Centre for the Study of Environmental Change, European
Parliament, Scientific and Technological Options Assessment, May 1996.

A. Rabl, Environmental Impacts and Costs: the Nuclear and the Fossil Fuel
Cycles, ExternE, Joule Programme JOU2-CT92-0236, European Commission,
June 1996.

101
LIFE CYCLE ANALYSIS AS BASIS FOR EVALUATING
ENVIRONMENTAL IMPACTS OF ENERGY PRODUCTION1

Edgar Furuholt
Statoil, Stavanger (Norway)

1. LCA methodology

To evaluate what LCA is suited for and what it is not suited for, it is
necessary to take a brief look at the methodology. The methodology and formal
framework has developed significantly during the last ten years. The most
commonly used frameworks are those given by SETAC and in ISO 14040. The
latter divides a Life Cycle Assessment (LCA) (see Figure 1) into four steps:

1. Goal and scope definition.


2. Inventory analysis.
3. Impact assessment.
4. Interpretation.
5. (Improvement).

2. Goal and scope definition

The problems of defining the goal and scope of a study should not be
underestimated. What is the purpose of the analysis? Do we for instance want to
compare two specific production chains for one identical product? Example:
Should we build a coal fired or a gas fired power station, each with a specific
technology and specified fuel sources, at a specific site? Or do we want to
compare average production chains in a country. The results of those two
analyses will most probably be quite different.

1. This article presents the views of the author and is not a position either of Statoil or the
International Gaz Union.

103
Figure 1. Life cycle assessment (LCA) framework

LCA framework
Direct applications

Goal & scope


definition
•Product development
and improvemnt
•Strategic planning
Inventory analysis Interpretation •Marketing
•Other

Impact assessment

Production chains will in most cases have several products that may be
spun off at different places in the chain. Natural gas is in most cases produced
together with oil that ends up in a large number of products. A gas power station
produces a lot of heat that may be used for district heating, in nearby industrial
plants etc. and the low temperature heat may be used for e.g. fish farming (see
Figure 2). How do we allocate the impacts in the production chain to each of the
products?

Figure 2. Energy chain

Fish farming
>100 % ?

200 C
Heating

Steam 80 %
Boiler

Fuel gas
Exhaust 500 C
Steam
turbine
60 %
Electricity
Gas Turbine to grid
45 %

104
Consider various gas power stations as an example, a simple gas turbine
power plant will have a typical efficiency of 45%. A combined cycle gas power
plant, where the turbine exhaust heat is used to produce steam which again
drives a steam turbine, may typically have an efficiency of 60%. The combined
cycle plant will then have a 25% lower CO2 emission per kWh. If the exhaust
heat instead is used to produce hot water for home or industrial use, the overall
efficiency may exceed 80%, but the electrical efficiency may drop to, say, 50%.
The CO2 emissions per kWh will then increase again by 20% if only the
electricity generation is considered, but decrease with 25% if the total energy
production is considered. Both results may be correct, depending on the purpose
of the LCA. If the low temperature water is used for fish farming, what is then
the functional unit of the total plant?

Some industrial plants have surplus energy, which may be used to generate
electricity for export to the grid. The electricity may then be regarded as a by-
product. How should the impacts of the plant be allocated to the electricity?

These simple examples serve to show that it is very important to be


completely clear on the purpose of an LCA, and also illustrates that results from
one LCA can not uncritically be used for other purposes.

3. Inventory analysis

Inventories for “traditional” parameters like resource consumption,


emissions, discharges, energy use, waste generation and treatment, land use,
noise etc. will normally be of good quality (high accuracy), at least for simple,
well-defined, production chains. It is more difficult to describe other parameters
like visual intrusions, aesthetical disturbances, changes in migratory patterns for
animals or birds etc. Non-regular phenomena, like accidents pose special
problems, particularly major accidents with low probability but severe
consequences.

With a view back to the first step of the analysis, goal and scope: How
representative is the inventory that is established? An example: The CO2
emissions for gas production and transport from the Norwegian North Sea
ranges between approx. 20 and 230 g CO2/Sm3 gas with an average of approx.
3
130 g/Sm , depending on which field the gas is coming from (Figure 3). This
corresponds to approx. 1-10% of the energy content in the gas, with an average
of approx. 5%. For one field, the value may vary by a factor of 3-4 over the
field’s lifetime. Inventories from two seemingly identical production chains may
therefore show large differences. Again, it is very important to be completely
clear on the purpose of the analysis. An inventory that is adequate, or even
absolutely correct, for one purpose may be completely misleading for another. In

105
fact: If you tell me the answer (e.g. energy consumption or emissions) that you
want (within reasonable limits) regarding oil and gas production, I will find a
production scenario in the Norwegian North Sea that will provide your results in
a way that is absolutely correct according to the LCA methodology. The same
possibilities for “manipulating” the inventory exist for many production chains,
implying that externalities based on LCA inventories must be treated with
appropriate care.

Figure 3. CO2 emissions from Norvegian oil and gas production


35 kg CO2/Sm3o.e.
( 0 – 100)
Oil to refinery
Emissions and
discharges
Gas to terminal

42 kg CO2/Sm3o.e.
( 0 – 120) 130 kg CO2/Sm3o.e.
( 20 – 230)

Gas injection Oil and gas from the reservoar

4. Impact assessment

The impact assessment is a much more complex and difficult step than the
inventory analysis. Contributions to local and regional effects cannot simply be
added throughout the production chain. NOx emissions for instance, which
contribute both to air quality/health impacts, acidification, eutrophication, ozone
formation and global warming, will have different impacts depending on the
geographical site of emission (see Figure 4). Both the total impact and the
relative contributions to these impact categories will be different depending on
the location. Where as air quality/health impacts will probably be the most
important category in urban areas, eutrophication and acidification will probably
be relatively more important in rural areas. It may therefore, for instance, be
sensible to spend more energy in a refinery, and thereby increase its (NOx)
emissions, to produce a gasoline quality that may lead to reduced (NOx)
emissions from cars driving in urban areas.

106
Figure 4. Impact assessment

Acidification

Eutrophication

NOx Ozone formation

Respiratory problems

Global warming Normalization

Such “relocation” of emissions may, however, lead to other effects that may
be hard to evaluate in an LCA. Example: we performed an LCA of conventional
gasoline and gasoline with MTBE some years ago. According to the “normal”
environmental parameters, the two gasoline qualities seemed to give comparable
total impacts but parts of the emissions were “relocated” from the roads to the
sites of the refinery and MTBE plant. A 10% increase in energy consumption,
and thereby emissions, to produce gasoline with MTBE, was offset by 0-5%
reduction in NOx, CO and HC emissions from the cars. What we did not foresee
and include, was possible negative impacts of MTBE on e.g. groundwater and
on people’s health that has come into focus in the last few years. Even if we had
known them, how could we have brought them into the analysis in comparison
with the other impacts to evaluate the total effects? If we should have tried to use
monetary values, how should we value people’s feeling of (lack of) well-being?

The ExternE study also clearly shows that equal emissions/loads may result
in widely different (monetary) impacts in different countries.

Impact assessment is therefore a troublesome task, even for products or


production chain where the main impact categories are similar.

When the impact categories are widely different, there is no objective way,
perhaps even not a scientifically valid way, to compare the total impacts of
different products or production chains. A number of methods for
“normalisation” have been proposed and are regularly applied in LCAs. Such
methods, including monetary valuation, may give valuable information when
interpreting the results from an LCA, but they may also be completely
misleading. The simple reason for this is that there is no unambiguous way of
comparing the effects like acid rain, climate change, peoples perception of noise,

107
visual impact, aesthetics, risks for major accidents, health impacts, security of
supply, risk for proliferation of radioactive material, peoples perception of risks
etc. All attempts to compare these will inevitably involve political, ethical and
emotional aspects as well as personal and societal dimensions and priorities will
have to be made (Figure 5). The way these priorities are made are obscured in an
analysis where the simple output is a monetary value and it will require a
significant effort and quite a lot of skill to follow the thoughts and assumptions
made by the analyst or developer of the method. These steps should therefore
rather be performed as an open, transparent political/public discussion than a
calculation in an LCA. A further complication is that even the comprehensive
ExternE project has not been able to value several important impact categories.
Other speakers will cover the challenge of setting monetary value on different
potential impacts in more detail.

Figure 5. No objective scale

Accident risk
Acidification Eutrophication
Noise Aesthetics Visual impact
Ozone formation
Security of supply Respiratory problems
Ozone depletion
Global warming Cancer risk Radioactive waste

Ethics
Politics
Emotions
Social
Personal
……

(Economic) value of all impacts

LCAs are therefore not suited for comparing impacts of widely different
electricity generation options like fossil, wind, hydropower, nuclear and solar and
therefore not suited as basis for political decisions regarding the balance between
these. They may, though, be suitable for comparing different transportation fuels.
Statements like “New research reveals the real costs of electricity in Europe”,
which have been used to describe the results of from the ExternE project, are
therefore simply not true. ExternE at best reveals a possible way of setting
monetary value to some of the impacts of electricity generation.

108
5. Interpretation

This final step of an LCA should evaluate the results of the analysis and all
choices made during the course of the analysis in terms of soundness and
robustness, and overall conclusions should be drawn.

As should be evident from the above discussions, an LCA that is designed to


give answers to specific questions, as posed in the goal and scoping phase, may
give anything from valid to misleading answers to other questions. The results
from an LCA made by an industrial company for evaluating different production
technologies before erecting a new factory, may therefore be unsuitable, even
misleading, for authorities when deciding taxes or other incitements for driving the
development in preference of certain products or production technologies.

The discussions and few examples above illustrate that LCAs have many
limitations and pitfalls. The results from LCA should handled with care and with
due consideration of its goal and scope. On the other hand, LCAs may give
valuable information in many decision processes if they are used for what they
are suited for and as one of several tools to clarify positive and negative effects
of a product, a process or activity.

According to our experience and opinion, LCAs may be suited for:

• Systematic description of resource consumption, environmental stressors


(emissions, discharges, waste, noise), possibly also accident risks and
consequences and to describe where (geographically) the environmental
burdens occur.

• Highlighting environmental challenges and to pinpoint where in the


production chain they can be found.

• Comparing impacts of products or production chains with similar impact


categories, like electricity production from gas, oil and coal.

• Comparing alternative technology options for a factory, including


technologies for reduction of emissions and discharges.

• Comparing alternative locations for a facility.

LCAs are not suited for:

• Comparing impacts of products or production chains with widely


different impact categories, like electricity production from fossil fuels
with e.g. hydropower, wind or nuclear.

109
• Internalisation of externalities for energy production and supply.

• Making decisions on taxes or other instruments to influence power


generation technologies unless used in combination with other tools and
evaluations.

• Industry evaluation of “green” taxes.

The work on externalities is important and interesting, but one should be very
conscious of the limitations of the available methods and that they will never give
objectively correct answers. The current state of the art regarding externalities of
energy production and products is far from giving a complete picture, although
significant progress has been made during the last decade. The results from studies
of externalities should therefore never be regarded as more than one of several
inputs in any decision process, whether in industry or in politics.

110
THE EXTERNAL COST OF THE NUCLEAR FUEL CYCLE

Caroline Schieber, Thierry Schneider


CEPN, France

Abstract

The external cost of the nuclear fuel cycle has been evaluated in the particular
context of France as part of the European Commission’s ExternE project. All the
steps in the fuel cycle which involve the use of cutting edge technology were taken
into consideration, from mining of uranium ores to waste disposal, via construction,
dismantling of nuclear power plants and the transport of radioactive materials. The
general methodology adopted in the study, known as the “Impact Pathway
Analysis”, is based on a sequence of evaluations from source terms to the potential
effects on man and the environment, and then to their monetary evaluation, using a
single framework devised for all the fuel cycles considered in the ExternE project.
The resulting external cost is in the range of 2 to 3 mEuro/kWh when no discount
rate is applied, and around 0.1 mEuro/kWh when a discount rate of 3% is
considered. Further developments have been made on the external cost of a nuclear
accident and on the integration of risk aversion in its evaluation. It appeared that the
external cost of a nuclear accident would be about 0.04 mEuro/kWh, instead of
0.002 mEuro/kWh without taking risk aversion into account.

111
1. Introduction

In the context of the Joule Programme of the European Commission, the


ExternE project has been implemented from 1991 to 1995 in order to assess the
external costs of various fuel cycles used in the production of electricity [1].
Within this project, the CEPN (Centre d’étude sur l’Évaluation de la Protection
dans le domaine Nucléaire) has completed the assessment concerning the nuclear
fuel cycle in the particular context of France [2]. All steps of the fuel cycle which
involve the use of cutting edge technology were taken into consideration, from
mining of uranium ore to waste disposal, via construction, dismantling of nuclear
power plants and the transport of radioactive materials. The general methodology
adopted in this study, known as the “Impact Pathway Analysis”, is based on a
sequence of evaluations from source terms to the potential effects on man and the
environment, and then to their monetary evaluation.

The aim of this paper is to present the main results of this study and to
discuss the meaning of the different indicators and assumptions adopted in the
evaluation of the external costs. As far as the nuclear fuel cycle is concerned, a
first set of questions arises from the risk assessment (especially the long term
and global impacts, as well as the integration of potential consequences
associated with severe reactor accident). Beyond this quantification of the
physical impacts on man and the environment, the evaluation of the external
costs associated with the nuclear fuel cycle also points out some questions on the
use of economic indicators (monetary value of life, discount rate, risk
aversion, …). Although the evaluation of the external costs implies the adoption
of various assumptions, it allows to put into perspective the health and
environmental effects of the different fuel cycles as far as their external costs
have been evaluated within a similar framework.

2. Methodology

2.1 Stages of the nuclear fuel cycle

The French nuclear fuel cycle was broken down into 8 separate stages (see
Table 1). Reference sites and 1990’s technology were chosen to represent the
total nuclear fuel cycle, as it exists today. In addition, the transportation of
material between the sites was considered. The facilities are assessed for routine
operation, except in the cases of electricity generation and transportation, where
accidental situations are evaluated. The impacts of construction and
decommissioning of a facility are included in the electricity generation stage. It
is important to stress that this methodology does not employ a worst case
scenario analysis, as is usually done for safety or regulatory compliance
assessments, but intends to evaluate the impacts expected from the operations. In

112
a few cases, however, when no reasonable alternative seemed possible,
conservative values were used.

Table 1. The different stages of the nuclear fuel cycle


considered in the French evaluation of the external cost

Stage of
Site Technology used
the fuel cycle
Mining and milling Lodève Underground and open pit mines
Conversion Malvesi and Pierrelatte Yellowcake conversion to UF6
Enrichment Pierrelatte Gaseous diffusion
Fuel fabrication Pierrelatte Conversion of UF6 to UO2 pellets
Belleville, Flamanville,
Electricity generation Saint-Alban, Nogent, 1 300 MWe PWR
Paluel
Reprocessing La Hague PUREX process
Aube Surface disposal
Waste disposal Auriat
Underground disposal
(hypothetical site)
Transportation Road and rail

The impact pathway approach requires an inventory and assessment of all


potential impacts, however, within the context of the ExternE project it has not
been possible to consider all of these. Therefore, only the most important
impacts, called priority impacts, have been included. Releases of radioactive
material to the environment, which potentially impact public health, were given
the highest priority. Occupational health impacts, from both radiological and
non-radiological causes, were the next priority, even though the extent to which
occupational health impacts can be considered as externalities has not been
addressed in this study.

2.2 Impact assessment

Assessment of the impacts was organized by the type of routine emissions:


atmospheric releases, liquid releases and solid wastes. The analysis of impacts of
releases from severe accidents involves additional complex issues, therefore, it
was treated as a distinct category. Health impacts to the workers – radiological and
non-radiological – were also accounted for separately.

The most important choices for the assessment of the nuclear fuel cycle
concern the definition of temporal and spatial boundaries. In this study a

113
distinction has been made between local impacts (0 to 100 km), regional impacts
(100 to 1 000 km), and global impacts (above 1 000 km) in the short term (first
year), the medium term (1 to 100 years), and the long term (100 to 100 000 years)
(see Table 2). The selection of a 100 000 year limit to the assessment was
arbitrary, however the most significant part of the impacts are included.

Table 2. Distribution of impacts from routine operation


of the nuclear fuel cycle

Local Regional Global

0-100 km 100-1 000 km >1 000 km


Non-radiological impacts
Short term
on workers
(<1 year)
Traffic accidents
85 3
Medium term Radiological impact Radiological impact Kr, H
14 129
(1-100 years) on workers and the public on the public C, I
Long term Radiological impact Radiological impact 14 129
C, I
(100-100 000 years) on the public on the public

The radiological health impacts for routine operations, and the number of
potential fatal cancers, non-fatal cancers, and severe hereditary effects in future
generations were estimated using recommendations of the International
Commission on Radiological Protection [3]. The non-radiological impacts of
accidental deaths and injuries for the general public, and the number of deaths,
working days lost and permanent disabilities for the workers were based on
published statistics. In accidental situations, additional impacts of the immediate
(deterministic) radiological impacts and the costs of the radiation protection
countermeasures were calculated using the COSYMA code [4].

The final stage of the impact pathway methodology is the economic


valuation of the impacts. The economists involved in the ExternE project set a
common value of a statistical life of 2.6 MEuros, based on a literature review of
contingent valuation studies, adopted for all the fuel cycle assessments in the
project [5]. Nevertheless, recently, lower values have been observed, especially
in a French contingent value concerning road accidents. In this survey, the value
of statistical life was about 1 MEuros [6]. Such a difference may largely modify
the result of the external costs of the nuclear fuel cycle, which is mainly due to
health effects. Furthermore, a working day lost and a permanent disability were
considered to be the equivalent of 65 Euros and 19 kEuros, respectively. These
values are based on those used by the French national health insurance system.
The value of 0.25 MEuros, adopted for a non-fatal cancer, was based on
information on direct and indirect costs from the United States of America. The
costs were calculated with an annual discount rate of 3%, adopted as the
reference value. Nevertheless, because of the large sensitivity of the results

114
according to the value of the discount rate, additional calculations were
performed for an annual discount rate of 10% as well as without discount rate.

2.3 Severe accidents

Accidents are one of the most controversial features of environmental


assessment of the nuclear fuel cycle. Within the scope of this project, this type of
assessment has been confined to the electricity generation stage and the
transportation of radioactive materials between sites. Although facilities at other
stages of the nuclear fuel cycle handle very large inventories of radioactive
material, their activities are generally believed to be of a lower risk. The
probabilistic assessment of the transportation of materials between all the fuel
cycle facilities includes risks from both conventional traffic accidents and
releases of radioactive material. These have been found to be relatively small.

At this time, there is no general consensus on a methodology to assess the


external costs of severe nuclear reactor accidents. In this project, a risk-based
approach has been adopted. Due to the complexity of the assessment and the
difficulty in finding facility-specific or generally accepted input data, the evaluation
that was completed provides indicative results for this type of methodology.

The source term considered in this study corresponds to a release of about


1% of the core (ST21). This source term is in the same order of magnitude as the
reference accident scenario used by the French national safety authorities. To
illustrate the sensitivity of the results the impacts of three other source terms are
presented. The largest can be considered as release that would occur after a core
melt accident with a total containment breach. The fraction of the core released,
based on a source term used in an international inter-comparison study, is about
10% of the core inventory. The smallest release can be considered to represent
the situation after a core melt accident where all the safety measures have
operated as planned and there is only leakage from the intact containment
(0.01% of the core inventory).

The probability of a core melt accident, based on a French assessment of a


major core melt accident at a 1 300 MW PWR reactor, is taken to be 1.0E-5 per
reactor.year. This is broadly consistent with other similar assessments based on
engineering fault tree analysis, although a wide range of estimates have been
proposed. The conditional probabilities of the large and small releases that
would occur after a core melt accident are taken from a US Nuclear Regulatory
Commission report, and are 0.19 for the three largest source terms and 0.81 for
the lowest.

115
3. Results

3.1 Doses

The total collective dose for all the stages of the fuel cycle, except for the
severe accident analysis, integrated for a time period of 100 000 years into the
future, is 13.1 man.Sv/TWh. A closer look shows that the total local collective
dose is about 0.22 man.Sv/TWh and the total regional collective dose is
0.33 man.Sv/TWh, leaving over 95% of the public dose due to the global
14 129
dispersion of certain radionuclides ( C, I). If the global doses are not included,
the occupational doses become a dominant contributor to the overall impacts
(about 40% of the doses received) (see Figures 1 and 2).

Figure 1. Distribution of the total collective dose associated with all stages of
the nuclear fuel cycle (without accident), integrated over 100 000 years
0.22
0.35
man.Sv/TWh
man.Sv/TWh 0.33
(1.7%)
(2.7%) man.Sv/TWh
(2.5%)

Total collective dose :


13.1 man.Sv/TWh

Occupational
Public local
12.20 Public regional
man.Sv/TWh Public global
(93.1%)

When all the categories of the doses are considered, the reprocessing stage,
with a collective dose of 10.3 man.Sv/TWh contributes the largest portion (79%)
of the total collective dose, followed by the electricity generation stage (18% of
the total). If the global collective doses are excluded, the reprocessing stage
diminishes in importance and is replaced by the electricity generation
(0.38 man.Sv/TWh) and the mining and milling (0.29 man.Sv/TWh) stages. The
doses from the enrichment stage are the least important. Tables 3 and 4 present
the order of magnitude of the doses associated with the different stages of the
fuel cycle, respectively for occupational and public exposures.

116
Figure 2. Distribution of the collective dose associated with all stages
of the nuclear fuel cycle (without accident), without global assessment
0.35
man.Sv/TWh
0.33
(38.5%)
man.Sv/TWh
(36.8 %)

Total collective dose :


0.9 man.Sv/TWh

0.22 Occupational
man.Sv/TWh
(24.7%) Public local
Public regional

Table 3. Order of magnitude of the occupational exposure

Occupational exposure (0-100 years)


Individual average Collective dose
Stage of the fuel cycle
(mSv/year) (man.mSv/TWh)
Mining and milling 2 to 5* 112
Conversion 2 2
Enrichment 2 <1
Fuel fabrication 7 6
Electricity generation 3 202**
Dismantling n.d. 22
Reprocessing 1 ~1
Transportation n.d. ~1
* Respectively open pit and underground mines.
** Average value for 1 300 MWe PWR.
n.d.: value not available.

In case of a severe reactor accident, an indicative total collective dose for


the population (for a radius of 3 000 km) for the four accident scenarios has been
estimated. The impact of the reference scenario ST21 (core melt with a 1% of
the core released) is a collective dose of about 58 000 man.Sv. For the other
scenarios considered, the expected risk (consequences x probability of
occurrence) varies between 0.001 and 0.08 man.Sv/TWh.(See Table 5)

117
Table 4. Order of magnitude of the public exposure

Public exposure (0-100 years)


(man.mSv/TWh)

Local Regional Global


Stage of the fuel cycle
(0-100 km) (100-1 000 km) (>1 000 km)
Mining and milling 83 90 <1
Conversion <1 <1 <1
Enrichment <1 <1 <1
Fuel fabrication <1 <1 <1
Electricity generation ~1 16 140
Dismantling <1 0 0
Reprocessing <1 84 480
Transportation ~1 0 0

Table 5. Expected collective doses for a major reactor accident


(ST21: reference scenario for France)

Source term Core melt Conditional Collective Collective dose x Risk*


(% of core probability probability dose probability man.Sv/TWh
released) (per (man.Sv) (man.Sv per
reactor.year) reactor.year)
ST2
1E-05 0.19 291 200 0.55 0.078
(10%)
ST21
1E-05 0.19 58 300 0.11 0.016
(1%)
ST22
1E-05 0.19 12 180 0.02 0.003
(0.1%)
ST23
1E-05 0.81 1 840 0.01 0.001
(0.01%)
* 7 TWh/reactor.year.

3.2 Human health impacts

3.2.1 Routine operations

The radiological health effects resulting from the normal operation of the
nuclear fuel cycle are directly proportional to the total collective doses. The
expected number of health effects were calculated assuming no lower threshold

118
for radiological impacts, using internationally accepted data from Publication 60
of the International Commission on Radiological Protection. The total number of
expected health impacts per TWh are: 0.65 fatal cancers, 1.57 non-fatal cancers,
and 0.13 severe hereditary effects. These results include the long-term global
dose assessment.

The number of estimated deaths for the European population due to the
routine annual operation of one additional 1 300 MWe PWR (about 7 TWh/y),
integrated over 100 000 years, would be less than 1 fatal cancer (0.1). This can
be compared to the approximate value of 800 000 fatal cancers reported in
Europe each year.

It is estimated that the production of 1 TWh will result in 0.02 deaths, 0.96
permanent disabilities and 296 working-days-lost (non-radiological health impacts)
in the work force for the nuclear industry. Worker accidents during the construction
and the decommissioning of the reactor are the most important contributors to these
values.

3.2.2 Accidental situations

The transportation of the radioactive materials between the different sites and
the transportation of the materials involved in the construction and the
decommissioning of the reactor result in traffic accidents involving the general
public. The number of non-radiological health impacts estimated are: 3E-4 deaths
and 1.7E-3 injuries per TWh. Assuming an incremental annual production of
7 TWh, less than 1 death (0.002) can be expected per year. This is insignificant
when compared to the nearly 10 000 traffic accident deaths that occur in France
each year. In accidental situations occurring during the transportation of hazardous
radioactive materials such as UF6, the toxicological health impacts estimated are
even smaller (2E-9 deaths/TWh and 7E-5 injuries/TWh).

The radiological health effects from reactor accidents can be divided into
two categories: the early health effects (deterministic effects) and the stochastic
effects as cancers or severe hereditary effects. For the four accident scenarios
considered in this study, only the two most severe accidents lead to deterministic
effects, but no early deaths are expected for the reference scenario ST21. For the
stochastic effects, as for normal operation, they are considered to be directly
proportional to the collective doses. Depending on the scenario, the number of
expected fatal cancers varies from 1E-4 to 3.9E-3 per TWh.

119
3.3 Monetary valuation

3.3.1 Routine operation

The sub-total of the cost presented for all the stages of the nuclear fuel
cycle is about 2.5 mEuros/kWh, if no discount rate is applied. When 3% and
10% discount rates are used, the cost is reduced to 0.1 and 0.05 mEuros/kWh,
respectively (see Figure 3). The current base load electricity generating costs in
France are on the order of 35-40 mEuros/kWh.

Figure 3. Distribution of the costs for the 0%, 3% and 10% discount rates

Sub-total = 2.52 mEuros/kWh Sub-total = 0.098 mEuros/kWh


No Discount Rate Discount Rate = 3%

5.5% 2% 2.5% 3%
8%

75% 14%

91%

Sub-total = 0.054 mEuros/kWh


Discount Rate = 10%
1% 2% 2%

95%

Public local Public regional Public global Occupational

The dominant contributor to the total cost is the reprocessing stage (76%),
followed by electricity generation (18%), when the 0% discount rate and the
global impact assessment are implemented. When the 3% discount rate is applied
(see Table 6), the construction of the reactor becomes the most important because

120
discounting does not reduce the costs of the very short-term impacts assessed
(40%). This is followed by mining and milling and electricity generation (19% and
17%, respectively). Six percent of the overall cost of the fuel cycle, at 0% discount
rate, is due to occupational health impacts. This proportion increases in
importance when a discount rate is applied (75% of sub-total for a discount rate of
3% and 95% of sub-total for a discount rate of 10%).

Table 6. Distribution of external costs of the nuclear fuel cycle with the
different stages of the nuclear fuel cycle, for public and workers

External cost
(routine operation, local and regional
areas, short and medium term impacts,
Stages of nuclear fuel cycle 3% discount rate)
(mEuros/kWh)

Public Workers
Mining and milling 5.24E-03 1.22E-02
Conversion 9.90E-08 4.50E-04
Enrichment 5.06E-08 7.47E-04
Fuel fabrication 2.66E-07 6.91E-04
Reactor construction 2.16E-04 3.70E-02
Electricity production 3.81E-03 1.02E-02
Dismantling 1.06E-02 6.41E-03
Reprocessing 1.21E-03 1.91E-03
LLW disposal 1.87E-12 2.46E-06
HLW disposal 0.00E+00 6.04E-09
Transportation 2.23E-04 2.80E-05
Total 2.14E-02 6.96E-02

The sensitivity to the discount rate used is due to the relatively large portion
of medium and long-term impacts associated with nuclear fuel cycle (see
Figure 4). For example, it can be seen that for no discount rate, the predominant
costs are due to the global assessment, however, if a 3% discount rate is used,
the short term, mostly occupational impacts dominate the final result. Another
example is the waste disposal stage where the relatively small costs disappear if
discount rates are applied. It has been for this reason that the use of this type of
impact pathway methodology for the assessment of waste disposal and global
impacts has been questioned.

121
Figure 4. Distribution of the costs for
the 0%, 3% and 10% discount rates with time and space (log scale)

Total: 2.52 mEuros/kWh Total: 9.83E-2 mEuros/kWh


No discount rate Discount rate 3 %

10 10

1E-1 1E-1

1E-3 1E-3

1E-5 1E-5

1E-7 1E-7

1E-9 1E-9
Global Short term Global
Short term
Regional Regional
Medium term Medium term
Local Local
Long term Long term

Total: 5.37E-2 mEuros/kWh


Discount rate 10 %

10

1E-1

1E-3

1E-5

1E-7

1E-9
Global
Short term
Regional
Medium term
Local
Long term

3.3.2 Accidental situations

The costs assessed for transportation impacts in general are extremely


small. The portion attributed to accidental conditions can be considered
insignificant due to the low probabilities and transportation packaging.

These results reported for the four accident scenarios are indicative of a
risk-based methodology. The reference scenario, considered to represent a core
melt accident followed by a release of 1% of the core, resulted in a
0.005 mEuros/kWh cost. The risk for the other scenarios varies between 0.02
and 0.0005 mEuros/kWh (see Table 7). The portion of these costs that might be
internalised by nuclear accident insurance has not been addressed.

122
Table 7. Results of accident analysis at a 1 300 MWe PWR, for 4 different
scenarios including public health effects and costs of countermeasures

Source term Cost x


Core melt
(approximate Conditional Total cost probability Cost*
probability
% of total probability (MEuros) (MEuros per (mEuros/kWh)
(per reactor.year)
core released) reactor.year)
ST2
1E-05 0.19 83252 0.158 0.023
(10%)
ST21
1E-05 0.19 17093 0.032 0.0046
(1%)
ST22
1E-05 0.19 3339 0.006 0.0009
(0.1%)
ST23
1E-05 0.81 431 0.003 0.0005
(0.01%)
* 7 TWh/reactor.year.

Further work must be done to evaluate other potential social impacts and
external costs such as, inter alia, public perception, risk aversion, disruption of
electricity supply, and decommissioning of the destroyed reactor. Besides the
difficulty in assessing these impacts and costs, the partition of externalised
versus internalised costs must also be evaluated.

4. Discussion

As mentioned above, the evaluation of the external costs for the nuclear fuel
cycle is confronted firstly to the difficulty to integrate over long periods of time
and large areas small individual levels of exposure (and thus their translation
into monetary terms) and secondly to the probabilities associated with the
accidental scenarios. The following paragraphs do not intend to propose
solutions for integrating these impacts into the external costs but rather to point
out some of the various dimensions of these specific impacts.

4.1 Nuclear accident

Calculation of the economic consequences of a severe reactor accident


points out the difficult question of the choice of a reference scenario for the
source term. The reference source term for the French safety analyses is focused
on about 1% of the core released after meltdown. Although there is a large
consensus within the scientists to consider a reference probability for the core
-5
meltdown around 10 per reactor.year, there is less analyses related to the
consequences of this meltdown: it is generally considered that in most of the

123
cases, the releases should be limited. Thus at this stage, simplified assumptions
were used in the ExternE project in order to aggregate the probabilities and
consequences. It is clear that this approach is limited as far as it does not
consider the risk perception associated with the potentiality that a severe
accident occurs as well as the indirect consequences on the economic activity.

4.1.1 Indirect costs

Thus, further to the classical evaluation performed for the health and
environmental consequences of a nuclear accident, it has to be considered that
there will be a decrease or an interruption of most of the economic activity
(essentially agricultural and industrial productions) in the affected territories
during a significant period of time. The importance of this interruption will
notably depends on the size of the accident. In terms of monetary indicators, this
disturbance of the economic activity will meanly induce a loss of value added
(this indicator corresponding to the different direct and indirect incomes of the
various “economic agents”). In order to derive the order of magnitude of the
indirect consequences, calculations have been performed using the COSYMA
code [4] and the value of life adopted in the ExternE project, and the following
results were derived [7]:

• The reference accident is supposed to induce an indirect cost which


represents about 10% of the regional gross domestic product during the
first two years.

• Similarly, it represents about 0.2% of the national gross domestic product.

• The indirect costs lead to an increase of 25% of the direct external costs
of a nuclear accident.

In fact, based on this calculation, one can consider a multiplying coefficient


of 1.25 to be applied to the direct external costs calculations in order to derive
the total external costs of the accident. This simplification seems to be
reasonable as soon as we are dealing with accident scenarios leading to
significant radioactive releases into the environment. In that case, the external
cost associated with the nuclear accident is 0.0057, for a reference source term
of 1%, instead of 0.0046 without taking into account the indirect costs.

4.1.2 Risk aversion

The main criticism of this approach is that there is a discrepancy between


the social acceptability of the risk and the average monetary value that in

124
principle corresponds to the compensation of the consequences for each
individual of the population affected by the accident. In fact, it appears that there
is a need to integrate risk perception – risk aversion within the calculation of this
external cost.

Some economic developments have been made, based on the expected utility
approach, in order to integrate risk aversion within the evaluation of the external
cost of the nuclear accident [8]. One of the interest of this approach is the
availability of experimental data concerning risk aversion coefficient. Although a
large range of values has been published for this coefficient, mainly based on the
analysis of financial risks, it seems reasonable to adopt a risk coefficient around 2
for the specific case of nuclear accident. This leads to estimate a multiplying
coefficient approximately equal to 20 to be applied to the external cost of a nuclear
accident corresponding to a release of about 1% of the core.

According to the previous results, the external cost of a nuclear accident


would be in the order of 0.092 mEuros/kWh, instead of 0.0046 mEuros per kWh
without taking risk aversion into account. In this case, it would represent about
3.6% of the external cost of the nuclear fuel cycle (with no discount rate, and
without accident).

4.2 Long-term global impacts

The major difficulty from the methodological point of view with regard to
the calculation of the external costs of the nuclear fuel cycle concerns the
3
evaluation of the impacts associated with releases of radionuclides such as H ,
14
C, 129I and 85Kr, and their translation into monetary terms. These releases induce
long term and global impacts because of their radioactive half-life or their
transfer into the environment. They were estimated using models and hypotheses
(migration into the environment, dose calculations, dose-response relationship,
constant population, etc.) that have been internationally agreed but which are
still being debated extensively.

4.2.1 Methodological considerations

The evaluation of long term and global impacts is raising various theoretical
and complex issues related to the validity of the quantitative assessment of what
could be the future risks but also to the ethical position we are adopting towards
future generations. However from a practical point of view, a responsible
attitude implies to use in the best way the available information we have about
the possible consequences of our present actions even if this information is just

125
reflecting limited knowledge taking into account the weakness of our present
instruments to assess far future consequences.

Although the concept of collective dose has not to be considered as the


“best indicator”, it allows to express the impact on populations in space and time
and provides additional information to the evaluation of individual exposures for
far future exposures. Nevertheless, one has to keep in mind the two types of
arguments that have been raised against the use of collective dose to assess the
impact far in the future:

• The aggregation on large population of extremely low individual doses


leads to significant collective doses.

• The existence of increasing uncertainties especially with time are


weakening the pertinence of estimation of impacts in the far future.

These criticisms are very valid, however any responsible attitude cannot
avoid taking into account the magnitude of the release of radionuclides in the
environment, the length of the period during which these radionuclides remain a
source of exposure and how wide they are dispersed geographically, i.e. how
large is the exposed population.

So far the use of the individual dose and collective dose concepts allow to
determine the order of magnitude of the long term and global impacts and to
assess if these impacts do not induce any problem in the future in terms of
individual risk or public health. In fact, there is a need to consider these various
indicators for evaluating the impacts because of the complexity of the problem.
In this context, their translation in monetary terms is difficult: any aggregation of
the complexity will reduce the dimension of the problem.

4.2.2 Order of magnitude of the impacts

Assuming a constant world population of 10 000 million people, the impacts


14 129
associated with the releases of long live radionuclides (mainly C and I) are at
the maximum of about 20 man-Sv per TWh, when integrating individual doses
over 100 000 years at the global scale. In terms of individual exposure, the average
-6
doses are bellow 10 mSv/year for the various releases of the French nuclear fuel
cycle associated with one year of production. For comparison, it should be kept in
14
mind that the individual dose associated with C which is naturally present into
-2
the environment is of the order of 1.2 10 mSv/year.

A second aspect concerns the mining residues. On the basis of an optimistic


assumption, that after the closure of the site, these residues are managed in order

126
to catch the radon emanations (over the natural background), the long term
impacts are negligible. In the absence of any specific treatment, the collective dose
associated with these residues is about of 200 man-Sv/TWh over a period of
10 000 years according to the estimates performed by UNSCEAR [9] for local and
regional populations. In the French evaluation, assumptions have been made on
the coverage of the old mines, and only the differential with the natural
background of radioactivity has been considered. for these reasons, the collective
dose considered in the study are far below the evaluation of UNSCEAR without
specific treatment (around 200 man.mSv/TWh).

Concerning the management of radioactive waste, the evaluation of collective


doses is rather limited according to the available studies in this field. In terms of
individual dose, a European study (Everest project [10]) provides some
evaluations for impacts associated with the storage of high level waste. For a
granite storage, the maximum individual doses are of the order of
-4
2 10 mSv/year after a period of 20 000 years in the case of the “normal
evolution” of the site which is designed to receive the waste associated with the
reprocessing of about 100 000 tons of spent fuel. As far as the intrusion scenarios
are concerned, the maximum individual dose is of about 2 mSv/year. For both
129
cases, the doses are associated with I. For low and intermediate waste disposal,
-3
the maximum individual doses for the public are of about 4 10 mSv/year (impact
3
associated with the releases of H) during the monitoring phase (300 years) and
-3
8 10 mSv/year above this period.

For the evaluation of the external costs, it appears that regardless of the
radionuclides and the period of time considered, the individual impact remains
insignificant (maximum annual individual dose estimated to be in the region of
-8
10 mSv/year per TWh). Integrating in space and time for the population as a
whole, the global external cost increases by at most 10 to 20% if a discount rate
of 3% is considered. This situation is quite different without discounting the
external costs of long-term impacts: in that case, the external costs should be
multiplied by a factor 10. In this context, the validity of an annual discount rate
and of the integration over very long period of time should be questioned.

5. Conclusion

From this assessment, it clearly appear that the main impacts are on workers
operating nuclear installations, with an individual level of exposure in the range
of a few millisieverts. As far as the public is concerned, the level of exposure for
the present generations (0-100 years) due to direct radioactive releases from
facilities into the environment is extremely low: a very small fraction is added to
natural background exposure and the corresponding individual risk can be
considered as negligible.

127
Concerning the long term and global consequences, the exposures of the
future generations remain negligible in terms of individual risk and one can state
that the total collective exposure that can be estimated by integration over a large
time period and the global population will never be a public health problem as the
worldwide collective exposure for a given generation (about 30 years) is at the
most a few man-sieverts.

A remaining issue is the potential for major accidents which may have
significant health and environmental consequences. The prevention of such
accidents calls for a high level of vigilance and on-going improvement of safety.

In this context, the monetary evaluation of these impacts points out some
issues for the economic theory. The development of contingent valuation
surveys during the last decade has improved the valuation process of these
impacts, but further analyses should be focused on the validity of aggregating
indicators using discounting procedure for very long term and global impacts as
well as indicators based on expected value in the case of major accident. Due to
the complexity of the health and environmental impacts, the monetary valuation
process has to be used with caution in order to reflect the various dimensions of
these impacts.

REFERENCES

[1] European Commission, Externalities of Fuel Cycle ExternE Project,


Workshop on the External Costs of Energy, CEC/OECD/IEA, Brussels
(Belgium), January 1995.

[2] M. Dreicer, V. Tort, H. Margerie, The External Cost of the Nuclear Fuel
Cycle: Implementation in France, CEPN Report R-238, August 1995.

[3] International Commission on Radiological Protection (ICRP), 1990,


Recommendations of the International Commission on Radiological
Protection, Publication 60, Annals of the ICRP, Pergamon Press, UK, 1991.

[4] J. Ehrhardt, J.A. Jones, An Outline of COSYMA, A New Program Package


for Accident Consequence Assessments, Nuclear Technology No. 94, 196-
203, 1991.

128
[5] A. Markandia, Externalities of Fuel Cycles “ExternE Project”, Report N° 9:
Economic valuation – An impact pathway approach, European Commission
DGXII, 1995.

[6] B. Desaigues, A. Rabl, Reference Values for Human Life: an Econometric


Analysis of a Contingent Valuation in France, In: N.G. Schwab Christe,
N.C. Soguel, Contingent Valuation, Transport Safety and the Value of Life,
Kluwer Academic Publishers, Boston/Dordrecht/London, pp. 85-112, 1995.

[7] T. Schneider, Integration of Indirect Costs Evaluation, CEPN-NTE-97/10,


April 1997.

[8] L. Eeckhoudt, C. Schieber, T. Schneider, Risk Version and the External


Cost of a Nuclear Accident, Journal of environmental Management, 58,
109-117, 2000.

[9] United Nations Scientific Committee on the Effects of Atomic Radiation


(UNSCEAR), Sources, Effects and Risks of Ionising Radiation,
UNSCEAR 1993 Report, United Nations, New-York, 1993.

[10] Institut de Protection et de Sûreté Nucléaire, Programme Européen


EVEREST : Étude comparative des résultats obtenus par l’IPSN concernant
les formations sédimentaires et granitiques, Rapport IPSN/DES; No. 278,
1996.

129
HYDROPOWER – INTERNALISED COSTS
AND EXTERNALISED BENEFITS

Frans H. Koch
IEA – Implementing Agreement for
Hydropower Technologies and Programmes, Ottawa (Canada)

Abstract

The benefits of hydropower consist of the minimal level of noxious and


greenhouse gas emissions, it’s energy security from political instability, and its
renewable, non-depletable nature. The costs of hydropower consist of negative
effects on the river ecosystem and of social changes in communities in the
vicinity of large projects. Public awareness of these costs has increased
dramatically during the past two decades, and new hydro projects will not get
approval unless adequate mitigation measures are taken to avoid, offset, or
compensate for adverse environmental and social effects. To a very large extent,
the hydropower industry has internalised what were previously social and
environmental externalities. However, hydropower operators do not receive any
compensation for the benefits, and to date their competitors (coal, natural gas,
oil) have not been required to internalise their adverse environmental
externalities. (emissions, depletion of supplies, and sometimes dependence on
imported primary energy sources). This creates an uneven playing field, and the
hydropower industry enthusiastically welcomes a discussion of this issue, and
eventually measures to rectify the situation.

The IEA Hydropower Agreement has completed a major international study on


the environmental and social impacts of hydropower, and one major component
of this study was a Life Cycle Assessment and comparison of all the most
important electricity generation technologies. The main conclusions were:

• The LCA methodology can give useful results for some specific
environmental parameters, and could probably be broadened to include
some social parameters.

131
• The LCA methodology is heavily dependent on the initial assumptions,
and the outcomes for a specific parameter can have a very wide range of
values, sometimes of two or three orders of magnitude.

• The LCA methodology does not currently take into account the
important issue of level of service, also known as ancillary services.

• In the case of hydropower, and probably in the case of other generation


technologies, a separate LCA analysis would be needed for each power
plant because the results can not be generalised to the industry as a whole.

• Some environmental parameters, such as biodiversity, and some social


parameters, such as esthetics or visual amenity are difficult or
impossible to quantify.

• The multiple impacts of any major infrastructure project are judged not
only by economic value systems (i.e. economic and financial rate of
return) but also by cultural value systems (health, education, public
safety, esthetics, heritage, social equity, conservation) which differ from
country to country.

• The LCA methodoly is especially strong in two parameters of greatest


importance to hydropower, noxious emissions and greenhouse gas
emissions. It could and should be used to internalise these externalities
either through imposing a cost on emitters, or paying a benefit to non-
emitters, or a combination of the two.

1. Introduction

Hydropower is characterised by its complete site specificity, each


hydropower project is located on a particular site with a particular topology on a
particular river in a particular climate zone and in a particular eco-system.
Hydropower projects can be found high in the mountains above the tree line and
down in lush valleys. They are found in arctic and sub-arctic regions as well as
tropical regions, in desert or semi-arid regions as well as regions with heavy
rainfall. They are found in completely un-inhabited areas and in densely
populated river valleys. It is not surprising that few generalisations can be made
about hydropower, and that for almost every generalisation that is made
exceptions can be found.

Aside from the high site specificity of hydropower, there are broad
categories of hydropower projects with very differing environmental impacts,

132
the main ones are run-of-the-river projects, projects with a large reservoir, and
multi-purpose projects (irrigation, flood control, navigation). Some
governments have also made distinctions between large hydro projects and
small ones, but hydropower professionals in general believe that on a per kWh
basis small hydro projects can have more environmental impacts than large
ones. A recent study by the German Ministry of the Environment supported this
conclusion, at least in the case of Germany.

In 1998, hydropower generated 2 643 TeraWatt hours of electricity (2 643 ×


9 1
10 kWh), which was 18.4% of the world’s total electricity production. If the
average wholesale price of electricity is taken to be about 3 cents US per kWh,
then the annual 1998 production would be worth $79 billion. It would have
2
required 1 586 million tons of coal to generate this same amount of electricity.

2. The IEA Hydropower Agreement Study of Social and Environmental


Impacts

The IEA Implementing Agreement for Hydropower Technologies and


Programmes, as it is officially known, conducted a major international study of
the Environmental and Social Impacts of Hydropower during the period 1995-
3
2000 . It involved six participating countries, who held 11 international
meetings and workshops. In all 112 experts from 16 countries, the World Bank,
and the World Commission on Dams have participated, and 29 professional
papers have been presented at meetings. One major component of this study
was a sub-task entitled: “Environmental and Health Impacts of Electricity
Generation – A Comparison of the Environmental Impacts of Hydropower with
those of Other Generation Technologies”. This sub-task was essentially an LCA
comparison of all the important electricity generation technologies: coal, oil,
natural gas, nuclear, biomass, hydropower, wind and solar. It will be called “the
LCA sub-task” in the remainder of this paper.

The LCA sub-task did a very extensive literature survey to collect from
various sources published data about the environmental impacts of the various
generation technologies. It did not do any original data acquisition of its own.
Although all six participating countries contributed to a greater or lesser extent,
the main work was done by two organisations, Vattenfall in Sweden and Hydro
Quebec in Canada. One result of their work is Table 1, below, which is a
reduced version of a larger table.

1. IEA, World Energy Outlook 2000.


2. The assumption is that it takes about 0.6 kg of coal to generate 1 kWh of
electricity.
3. Available on the web-site http:\\www.ieahydro.org.

133
Table 1. Emissions produced by 1 kWh of electricity
based on life cycle analysis

Generation option Greenhouse SO2 NOx emissions NMVOC Particulate


gas emissions emissions milligram milligram matter
gm equiv milligram /kWh /kWh milligram
CO2/kWh /kWh /kWh

Hydropower 2-48 5-60 3-42 0 5


Coal – modern 790-1 182 700-32 321+ 700–5 273+ 18-29 30-663+
plant
Nuclear 2–59 3-50 2-100 0 2
Natural Gas 389–511 4-15 000+4 13+-1500 72-164 1-10+
(combined cycle)
Biomass forestry 15–101 12-140 701-1950 0 217-320
waste combustion
Wind 7–124 21-87 14-50 0 5-35
Solar photovoltaic 13–731 24-490 16-340 70 12-190

The table clearly shows that the amount of noxious emissions (SO2, NOx,
NMVOC – non methane volatile organic compounds, and particulate matter)
and greenhouse gas emissions are much less for hydropower, nuclear, and wind,
than they are for fossil fuels (coal and natural gas). This fact has been widely
known for a long time, and in a qualitative sense it is nothing new. However,
being able to quantify these results by means of an LCA is an important step
forward, and should, in turn, enable policy makers to quantify the taxes or other
policy instruments that they may use to internalise these externalities.

3. Extension of the LCA approach

Further research work to expand the LCA approach to issues of energy


security would certainly be welcomed by the hydropower industry (and no
doubt by the other renewable energy industries as well). Although many rivers
cross international boundaries, and there are some international hydro power
projects, most hydro is a national resource and not dependent on political
stability of energy exporting countries. However, the amount of hydroelectricity
available in a given year is dependent on the amount of rainfall, and can easily

4. The sulphur content of natural gas when it comes out of the ground can have a wide
range of values, when the hydrogen sulphide content is more than 1%, the gas is
usually known as “sour gas”. Normally, almost all of the sulphur is removed from the
gas and sequestered as solid sulphur before the gas is used to generate electricity. Only
in the exceptional case when the hydrogen sulphide is burned would the high values of
SO2 emissions occur.

134
vary by more than 20% from year to year. Currently, reservoir levels in most of
Brazil and in the US North-West are low, and the electricity produced is less
than in average years. Hence energy security depends not only on political
factors, but also on the weather.

Similarly, work to expand the LCA approach to issues of energy depletion


would also favour all forms of renewable energy. Hydropower is
quintessentially a long-term investment, some plants have now been in
continuous operation for more than 100 years. Current economic thinking and
net present value calculations give little importance to cost or benefit streams
that occur more than 15 or 20 years from the present. Yet the hydropower
industry abounds with examples of communities or utilities which constructed
hydro plants thirty or forty years ago, and made little profit while the debt was
being amortised. However, they now own money spinning assets, which
unobtrusively and reliably produce thousands or millions of dollars of revenue
year after year. A net present value calculation thirty or forty years ago when
the decision was made to build such plants would have rated them as relatively
poor investments compared to others which might have had a higher internal
rate of return. Some of the current economic theories do not appear to give
adequate guidance when dealing with the long term future, whether it is long
term benefit streams or long term costs, such as depletion of fossil energy
resources. An extension of the LCA methodology to deal with these long-term
issues would be an important step forward.

4. Hydropower – Internalised costs and externalised benefits

The past two decades have seen a paradigm shift not only in environmental
issues, but also with respect to the rights of individuals, small groups, and local
communities affected by large infrastructure projects. The hydropower industry
has changed accordingly, and the way hydro projects are done today is very
different from the way they were done 20 years ago. Full environmental
assessments are done, and extensive consultations are held with affected groups
and communities, in a broad participatory decision making process. Most (but not
all) adverse impacts of a project are established during the planning stage, and the
design of the project will avoid impacts if possible, or include mitigation
measures, or provide for compensation. Modern hydro projects have to pay for
most of their social and environmental impacts, otherwise they will not get
government approval to proceed. In some cases reserve funds are set aside to pay
for impacts which are discovered once the project is completed and in operation.
It is not uncommon for hydro projects to be given the stewardship responsibilities
for an entire river basin and to have to pay for the impacts of other polluters. For
example, the Bonneville Power Administration in the North Western USA pays
not only for the environmental impacts of the hydropower industry, but also for

135
those due to forestry, agricultural run-off, and other industries.5 In most
jurisdictions, the hydropower industry has fully internalised the cost of adverse
environmental and social impacts, and in some case it has gone beyond.

In most countries, the main competitors of hydro power, i.e. coal and
natural gas generation, do not have to pay for their environmental impacts, and
especially their emissions of noxious gases and greenhouse gases. This could be
corrected by requiring them to pay for their emissions, i.e. internalising their
externalities, or by paying a bonus to renewable energies for not emitting, or by
a combination of the two. Creating a level playing field is important for the
hydropower industry, and the LCA methodology could make a very useful
contribution in this respect.

5. Experience with the LCA approach

The LCA sub-task has documented various aspects of the LCA approach,
which need to be taken into account in decision making.

5.1 Wide range of values

Firstly, the numbers in Table 1 give a very wide range of values, up to three
orders of magnitude in some cases. This is due to the specific site in some cases,
and due to the variety of assumptions that can be made about the materials and
manufacturing processes going into a project. Cement can be made in gas fired
kilns or coal fired kilns, the gas and coal can have a high sulphur content or a low
sulphur content. The electricity used in manufacturing processes can come from
thermal sources or renewable sources. All these assumptions have a determining
effect on the outcome of the LCA. The emissions of solar voltaic are a good
illustration, because a lot of electricity is required to manufacture a photovoltaic
cell. If this electricity is derived from coal, the greenhouse gas emissions during
the manufacture of the photovoltaic cell can be of the same order of magnitude as
natural gas combined cycle generation.

The range of values for a generation technology as a whole is so wide that


it does not provide very useful numbers for policy making. Instead, a specific
project will have to be selected first, and the LCA calculations made for that
project. Even then, it may not be known in advance from which country the
steel and cement will be bought, and the resulting calculations could have

5. Personal communication, the Bonneville Power Administration has a budget of


about US $ 400 million per year for environmental improvements in its service
area.

136
widely different answers as a result. If the LCA approach becomes more widely
used, the LCA parameters for all the main materials will be more readily
available, and with the use of computers the overall LCA for construction of a
plant can be calculated more easily.

The operation of a plant will also cause variations in environmental


impacts. The operation of a hydropower reservoir can make significant
differences both in the reservoir itself and downstream. The effects of a coal or
natural gas plant will depend on the sulphur content and other characteristics of
the fuel that is being used at any given time. These variations during operation
give rise to a range of values, even for a specific plant, and they are more
difficult to handle from the regulatory point of view.

5.2 Level of service

The consumer expects electricity to be available on demand, when he or


she turns on a switch the light should go on. Since electricity itself can not be
economically stored, utilities face a major challenge in continually adjusting the
amount generated to be equal to the amount demanded. Some generation
technologies, such as hydro or diesel, can easily adjust the amount of electricity
produced, whereas others such as coal or nuclear can only operate economically
at a constant output level. Yet others, such as solar or wind are subject to
interruption, either because the sun doesn’t shine or the wind doesn’t blow. The
ability of a generation technology to easily match the amount produced to the
amount demanded by consumers is known as level of service. In the power
industry the words ancillary services are used. In privatised electricity markets,
the exact market rules and the generation mix in the market area determine
whether, and to what extent, the ancillary services are compensated. In some
cases these market rules are satisfactory, and in others less so. The main point is
that economic mechanisms do exist to deal with the differing levels of service.
The LCA methodology does not in its present form include the issue of level of
service, and it is probably not necessary to do so because of the existence of
these other economic tools. However, the interpretation of LCA results should
always be done with the caution in mind that differences in levels of service are
not reflected.

5.3 Quantification of some environmental and social parameters

Some environmental parameters, such as biodiversity, and some social


parameters, such as esthetics or visual amenity are difficult or impossible to
quantify. For the hydropower industry, the biodiversity issue is important,
especially as related to fisheries of migratory species (salmon, trout), and to the

137
creation of new wetlands downstream or the inundation of existing wetlands by
a reservoir. The question of esthetics and visual amenity can be quite
controversial, some will be very happy with a beautiful new lake which can be
used for recreational and other purposes, whereas others will be unhappy
because a certain area of land is flooded. In many cases new transmission lines
have to be constructed for hydro projects, and this may detract from the visual
amenity of a landscape. The fact that these environmental and social parameters
are difficult or impossible to quantify means that an LCA analysis can not tell
the complete story about the desirability or otherwise of a proposed project.
Consequently, the LCA may be helpful in internalising some of the costs, but it
can not internalise all of the important effects.

5.4 Cultural value systems

The multiple impacts of any major infrastructure project are judged not
only by economic value systems (i.e. economic and financial rate of return) but
also by cultural value systems (health, education, public safety, esthetics,
heritage, social equity, conservation) which differ from country to country.

An LCA analysis may provide numerical values for several environmental


and social parameters, and it may then be possible to attach a monetary value to
each numeric environmental parameter, and then to internalise the cost. Once
each parameter has been translated into a monetary value, it becomes easy to
compare and manipulate them. One effect will cost 0.5 cent per kWh, another
0.1 cent, a third 1.5 cent, etc. Decisions about trade offs, cost of mitigation
measures, taxes or surcharges, etc. become straightforward.

When dealing with cultural value systems there is no quantitative


methodology for comparing either the costs or the benefits of a project against
each other. Bringing electricity to a town or village for the first time may start a
whole chain of events leading to increasing prosperity: light in the evening may
stimulate adult education, refrigeration allows for better storage of medicines
and food, craftsmen can increase their productivity with electric tools, water
may be delivered by electric pumps, etc. To do this, it may be necessary to
construct a coal fired generation plant, to build a transmission line through a
scenic landscape, and over the years a few unfortunate persons in a large
population may be electrocuted. The importance attached to each of these
factors, and the ultimate decisions made, will be based on cultural value
systems, and will differ from country to country. Even if some of these
parameters can be quantified, for example the increased literacy rate and the
accidental electrocution rate, they cannot be compared against each other. Even
if the LCA methodology enables a number to be attached to an environmental

138
or social factor, the next step of attaching a monetary value to that number may
not possible, and then the internalisation of that factor is also not possible.

6. Conclusion

The LCA methodology is especially strong in two parameters of greatest


importance to hydropower, noxious emissions and greenhouse gas emissions. It
could and should be used to internalise these externalities either through
imposing a cost on emitters, or paying a benefit to non-emitters, or a
combination of the two. If the LCA methodology can be extended to other
issues such as security of energy supply and depletion/renewability, the
hydropower industry would join the other members of the renewable energy
community in supporting this wider application.

At the same time, there are limits and challenges facing LCA methodology
and the process of going from a quantified environmental parameter to a
monetary value. Some important environmental and social parameters are
difficult or impossible to quantify. Some other parameters may be quantifiable,
but the next step of going from an environmental quantity to a monetary value is
difficult or impossible. Ultimately, decisions about approval of projects or
internalisation of their costs to society are based not only on economic values, but
also on cultural values. These cultural values differ from country to country and
are not easily incorporated into a quantitative framework.

The IEA Hydropower Agreement welcomes and encourages all efforts to


more fully internalise the costs to society of electricity generation. These efforts
should apply to all producers of electricity, at the present time the hydropower
industry has internalised its costs and some of its competitors have not.

139
LIFE CYCLE ASSESSMENT OF RENEWABLES :
PRESENT ISSUES, FUTURE OUTLOOK AND
IMPLICATIONS FOR THE CALCULATION OF EXTERNAL COSTS

Paolo Frankl
Ecobilancio Italia/Ambiente Italia, Roma (Italy)

Abstract

In principle, Life Cycle Assessment (LCA) is certainly appropriate for


estimating external costs of renewables, since major environmental impacts of
the latter are generated in phases of the life cycle other than use. In practice
however, several issues still remain. They are related to the availability and
quality of Life Cycle Inventory (LCI) data, to the fast technological development
of renewable energy technologies (RET), to the existence of many different
applications of the latter and to a strong dependency on local conditions.
Moreover, a “static” picture of present technologies is not enough for policy
indications. Therefore some kind of dynamic LCA is needed. These LCA issues
are reflected in the calculation of external costs.

First, the paper discusses these issues on the examples of two main technologies,
namely Photovoltaics (PV) and wind. Second, it discusses the results of ExternE
for these two specific technologies and gives an outlook for the future. Future
needs for a better use of LCA as a support tool for the calculation of external
costs are identified. Finally, a new research project funded by the European
Commission focused on LCI of renewables is briefly introduced and presented.

141
1. Introduction

In principle, Life Cycle Assessment (LCA) is certainly appropriate for


estimating external costs of renewables, since major environmental impacts of the
latter are generated in phases of the life cycle other than use. For example, solar
energy technologies have zero emissions during the use phase; a significant share
of the environmental impacts of wind is generated during the production of the
wind turbines. The CO2 emissions of biomass energy plants are compensated by
the CO2 absorption during the tree growth. All this justifies the use of a life cycle
approach in general and of LCA in particular as a supporting tool for the
estimation of external costs of renewable energy technologies.

However, in practice there are several open issues. Some main problems are
listed here:

• Availability and quality of LCA data. Many studies of LCA of


renewables have been carried out in the past. However, large
differences in results are reported in literature. This is discussed more in
detail later on for the case of photovoltaics (PV).

• Some renewable energy technologies are under fast technological


development (e.g. PV). Therefore, a “static picture” of present
technologies is not enough. Some kind of dynamic LCA and of
technological forecasting are needed in order to provide useful
recommendations for policy.

• Some RET have several different application scenarios. For instance,


PV can be used in independent Solar Home Systems (SHS) with
electric battery storage, in open-field power plants or integrated in
buildings. In the latter case, they might play a multi-functional role, as
they substitute building cladding materials, the may act as sun-shading
(and thus energy-saving) systems. Finally, they might also be used as
mini-hybrid systems with recovery of heat.

• All renewables strongly depend on local resources. Solar technologies


depend on the local solar irradiation, wind and biomass depend on the
local specific resource.

• As shown later, significant life cycle impacts of PV and wind are


generated during the use of (conventional) electricity for their
production. This implies that LCI results are strongly dependent on the
local electricity mix used for calculations. This rather reflects the

142
present technological status of a National system than of the technology
itself.

All these aspects obviously strongly influence the final result in terms of
external costs per produced kWh. The issues for PV and wind and their
implications for the calculation of external costs are discussed in the next
paragraphs.

2. Photovoltaics

2.1 Introduction

Several LCI or LCA studies of PV systems have been published in the


period 1989-1998 in Germany, Switzerland, Netherlands, UK, US, Australia,
Japan, Italy [1,3,6,10,12-15]. Large differences in results emerge from these
studies, which were not immediately explainable at first glance. In 1997, an
International Workshop entitled “Environmental Aspects of PV Power Systems”
was organised at the Dept. of Science, Technology and Society of the University
of Utrecht within the activities of the International Energy Agency
Implementation Agreement on Photovoltaics, (IEA PVPS Task 1 – see for
instance [2]. The goal of the workshop was precisely to collect most of the up-
to-date LCA information on PV systems and to try to understand such important
differences arising from different studies. Another international workshop
followed in Keystone, Colorado, the year after, organised by the US Brookhaven
National Laboratory. As an outcome of these workshops and of their own
research activities, several authors have reviewed and discussed the energy
requirements for the production, the energy payback time (EPBT) and the
contribution for CO2 mitigation of PV systems [4,7,5]. These authors have
reviewed the main literature and have identified min. and max. values for
primary energy requirements for the production of PV systems. As a
consequence, they have been able to determine min. and max. estimates for
EPBT, for energy yield ratios (EYR) and for specific CO2 emissions over the life
cycle of PV systems. Moreover, they have highlighted the importance of the so-
1
called “Balance of System” (BOS), in particular for future PV systems. Some of
2
their results are summarised and discussed in the next sections.

1. The BOS is the ensemble of mechanical structures and electrical devices other than the
module necessary for the operation of PV systems (e.g. supporting structures, electric
inverters, etc.).
2. This part is mostly taken by earlier publications by the author [Frankl & Gamberale,
1998, Alsema, Frankl & Kato, 1998].

143
2.2 Energy requirements for the production of PV systems

2.2.1 Crystalline silicon

The study by [4] points out that the published estimates (as of 1997) for the
energy requirement of present-day crystalline silicon vary considerably:
2
between: 4 200-11 600 MJ/m for multi-crystalline silicon (mc-Si); and between
2
6 000-13 900 MJ/m for single-crystalline silicon (sc-Si). Partly, these
differences can be explained by different assumptions for process parameters
like wafer thickness and wafering losses. However, the most important source of
differences is the energy requirement estimation for the silicon feedstock used to
produce PV wafers. Currently the majority of PV cells are made from off-spec
silicon that is rejected by the micro-electronics industry. As a matter of fact, the
major source of uncertainty is the preparation of silicon feedstock from
electronic industry scraps, involving two crystallisation steps. The present
manufacturing energy requirement very strongly depends on:

• The allocation criteria for the primary crystallisation step.

• The silicon content of the cell.

• The specific energy consumption rates for silicon purification.

This situation is depicted in Figure 1, which shows the primary energy


requirements for “present” (1997) mc-Si PV systems highlighting the large
difference between the max. estimate (mc-Si “high”) and the min. one (mc-Si
“low”. For comparison, also the energy requirements for amorphous silicon (a-Si)
are reported. Moreover, the results for the BOS for three different applications (PV
plant in the open field, roof integration and facade integration) are also shown. The
latter figures on a-Si and BOS are discussed in the next sections.

In the future the introduction of solar-grade silicon preparation processes


2
might significantly reduce energy requirements down to 2 600 MJ/m and make
the discussion about one or two crystallisation steps obsolete [4]. The
significantly lower expected energy consumption values will mainly be caused
by three factors, namely: i) a much higher silicon mass efficiency (that is a much
better use of silicon feedstock input per kWp output, ii) internal recycling and
iii) less specific energy consuming processes (i.e. faster Czochralsky and/or
directional solidification processes).

144
Figure 1. Primary energy requirements for silicon PV systems (1997)

16000 inverter (3.3 kW)

mc-Si high mc-Si low a-Si BOS

14000 module frame

a-Si equipment manuf.


12000 a-Si overhead operations

a-Si cell/module processing


10000
MJ/m2 module

a-Si substrate+encaps. mat.

a-Si cell material


8000
module assembly

cell processing
6000
wafering

Crystallization & contouring #2


4000
Crystallization & contouring #1

silicon purification
2000
mg silicon production

0
PV roof facade PV roof facade PV roof facade
plant plant plant

2.2.2 Amorphous silicon

The differences in published estimates for the manufacturing of amorphous


2
silicon modules (710-1 980 MJ/m ) can be explained by the choice of substrate
and/or encapsulation materials, and by whether the overhead auxiliary energy
use and energy consumption for equipment manufacturing are taken into account
or not. The cell material itself accounts for only a few percent of the total energy
requirements. The best estimate for present primary energy requirement for
2
amorphous silicon manufacturing is around 1 200 MJ/m [7]. Assuming a 6%
module stabilized efficiency, this corresponds to specific energy requirement of
20 MJ/Wp, which is significantly lower than the one of present crystalline silicon
modules (35-96 MJ/Wp for mc-Si). However, lower efficiencies and thus higher
BOS requirements can cancel out this advantage, at present and in the future.
The potential for improvement is lower than for crystalline silicon modules
(max. 30% energy requirement reduction) with current encapsulation materials.

2.3 The importance of BOS

P. Frankl in [6] has carried out a detailed analysis of building-integrated PV


systems and highlighted the importance of taking into account the BOS in LCI
calculations. For the comparison of PV systems two major categories are
identified, namely “conventional” installations (array field PV power plants),
and PV systems in buildings. The latter can be further classified into sub-

145
categories, corresponding to the part of the building on which the PV system is
applied (terrace or flat rooftop, tilted roof, facade, etc.). Furthermore, the
classification depends on whether the PV system is mounted on existing
structures (retro-fit systems) or designed together with a new building
(integrated installations). Finally, integrated hybrid systems with heat recovery
are also considered. Mean European values for the energy requirements and
emissions related to the production of materials are used for calculations.

The study has taken into account several applications on rooftops and
building facades. It also has included for comparison the analysis of a large PV
plant in the open field, namely the 3.3 MW power plant in Serre, Italy.

The results show that the primary energy content of a PV power plant is in
2
the range of 1 900 MJ/m . Whi1e most of the systems in buildings have a total
2
primary energy content of around 600 MJ/m . The high value for power plants is
caused by the high amount of concrete and steel needed for this kind of installation
in open fields. It is not expected to drop significantly in the future. On the contrary
PV applications in architecture profit from the bui1ding structures. The analysis
also has shown that in the future the primary energy needed to install building-
2
integrated systems might well further drop down to around 400 MJ/m for tilted
2
roofs and 200 MJ/m for facades. These improvements can be obtained by
reducing absolute quantities of materials and/or using large fractions of recycled
materials (especially in the case of aluminium). It is worth reminding that the
actual roof and facade systems analysed were almost the first pilot cases of PV
building integration in several countries. Therefore, it might be assumed that the
use of materials has not been optimised. Furthermore, it is worth noticing the
significant contribution of module frames in present-day systems. Its wide range
2
of energy content (300-770 MJ/m ) is due to large differences in the amount of
a1uminium used for the frames. In any case, PV modules are expected to be
frameless for all future applications.

2.3.1 Future installations – Possible optimisations


In the future, PV installations in buildings will likely be designed taking
into account the full life-cycle of materials. This is necessary for an energy-
conscious, energy efficient and environmentally sound design of the systems.
Two approaches can be followed, namely: to minimise absolute quantities of
materials and to use a large fraction of recycled, secondary materials. Figure 2
shows the possible primary energy content of future optimised PV systems. The
scenario depicted is characterised by the following assumptions:

• Future installations will contain 80% of secondary aluminium. This


strongly decreases energy consumption for most PV systems in buildings.

146
• Light concrete supporting structures will likely be used for PV systems
on flat roofs, both for economic reasons and for the simplicity of
installation and maintenance.

• An advanced type of clay will be used for PV tiles, which allows


energy consumption to be reduced by about 30% [6].

If all the above mentioned factors are taken into account, the comparison
between the BOS energy content of PV plants and PV systems in buildings
becomes radically favourable to the latter, as clearly illustrated in Figure 2.

Figure 2. Possible future BOS primary energy content


of optimised PV systems

2000

copper
1800

1600 advanced
clay
1400 PVC

1200
MJth/m2

reinforced
concrete
1000
light
800 concrete

secundary
600 aluminium

400 primary
aluminium
200 steel

0
Integrated facade

PV cladding system
Flat roof - Light concrete

Retrofit facade

PV tile
Integrated tilted roof
PV plant

Retrofit tilted roof

2.4 Energy profiles of present and future silicon PV systems

2.4.1 Present systems (1997)

Figure 1 shows the primary energy requirements for present PV systems.


As a consequence of the high energy requirement for crystalline silicon module
manufacturing, the contribution of BOS is of minor importance, at least in the
case mc-Si “high”. As a matter of fact, in this case the Energy Pay-Back Time

147
(EPBT)3 is slightly higher than eight years, even if the system is installed in a
4
place with a relatively high sun radiation, such as Central Italy. Because of the
large contribution of PV modules, the installation of PV systems in buildings
reduces the EPBT only to a limited extent (max. 18% for roofs). Facades show
even worse results because of the bad exposure to the sun at these latitudes. The
most effective PV system seems to be the simple installation on flat roofs [7].

However, in the case of mc-Si “low” and a-Si, the contribution of the BOS
is proportionally higher. Therefore, the benefits of the integration in buildings
are more significant. Even the contribution of the aluminium module frames is
not negligible.

The proportional importance of BOS will increase further with future PV


modules and is described in more detail in the next paragraph.

2.4.2 Future prospects

In future, the manufacturing of crystalline silicon cells will require


significantly less energy. Whatever the specific technology (single- and/or multi-
crystalline silicon derived from electronic industry, or solar-grade silicon), the
production chain will be optimised for solar energy cell manufacturing. A much
smaller amount of silicon feedstock will be required to produce a cell. The cell
and module efficiencies will also increase. Also the technology of amorphous
silicon is expected to improve significantly. Table 1 summarises the expected
technological evolution of silicon modules adapted from P. Frankl [6] and
E. Alsema et al. [4].

Figure 3 shows the EPBT of future multi-crystalline silicon PV systems.


Results are subdivided according to the various manufacturing steps of crystalline
silicon PV systems, namely the preparation of high-purity silicon feedstock, the
cutting of silicon ingots into wafers, the manufacturing of cells, the assembling of
modules, and the BOS. Moreover, the contribution of process electricity and
primary energy content of materials are distinguished.

3. The EPBT is the time needed for the PV system to supply the amount of energy
consumed for its production. It is defined as: EPBT (years) = Consumed primary
energy for system production/Annual primary energy produced by the system.
4. Other parameters used for calculations are: i) PV plant electric BOS efficiency: 85%;
ii) efficiency of Italian electricity production mix: 39,1%; grid distribution losses: 7%;
iii) for integrated systems, the primary energy content of the building materials substituted
by the PV components have been subtracted from the BOS primary energy content.

148
Table 1. Technological parameters of present and
future silicon PV modules

Silicon PV technologies

Multi-crystalline Single-crystalline Amorphous


Present Future Present Future Present Future
Cell efficiency 14% 16% 15.5% 18% – –
Module efficiency 12.1% 14.5% 12.7% 14.8% 6% 10%
Primary energy content 4 200- 6000-
for module manufacturing 11 600 2 600 13 900 3 100 1 200 840
2
(MJ/m )
Module lifetime (years) 25 30 25 30 10 15

As far as this is concerned it is worth noticing the large contribution of


(conventional) electricity consumption for the production and purification of the
silicon feedstock..

As a consequence of manufacturing and efficiency improvements, the


expected EPBT of such “optimised” power plant is reduced by more than a
factor three (from 8 down to 2.3 years) with respect to present power plants.

Moreover, as already mentioned the BOS plays a more important role in the
total energy balance. This means that the integration in buildings gives
proportionally more benefits than today. The EPBT of a fully integrated future
mc-Si PV roof system is expected to be about 40% smaller than that of a future
PV plant.

Moreover, the EPBT is further strongly reduced if heat recovery is taken


into account. In integrated systems, at least part of the heat dissipated by the PV
panels can be recovered by means of an air channel between the back-plates of
the modules and the roof (or facade) itself. This air flow has a double effect:
first, it allows the warm air to be used in the building for air conditioning and/or
pre-heating of water; second, it cools down the cells, thus increasing their
5
efficiency. In this case, the thermal energy recovery in tilted roof can reduce the
EPBT by almost a factor 3 with respect to a PV power plant. As a matter of fact
the expected EPBT of an integrated tilted roof with heat recovery is lower than
6
10 months. It is also worth noticing that the PV facades become interesting

5. In this study an annual mean value of 2 kWhth recovered heat per kWhel produced by
the PV system is assumed
6. To calculate the corresponding primary energy, the substituted heat has been supposed
to be produced by methane boilers.

149
when equipped with a heat recovery system (Figure 3). However, given the
difficulties to effectively recover and use the thermal energy throughout the
whole year, these results have to be interpreted with care.

Figure 3. Expected energy pay-back times for


future optimised multi-crystalline silicon PV systems
2
(mean annual insolation: 1 700 kWh/m year on a 30° tilted,
south-oriented surface; cell efficiency: 16.0% ; module efficiency : 14.5%)

2.50

2.00

BOS materials en.


1.50 cont.

Module electricity
EPBT(years)

1.00
Module materials
en. cont.

0.50 Cell electricity

Wafer electricity
0.00
PV power plant

flat roof

tilted roof

facade

tilted roof

Alukobond facade

tilted roof

Alukobond facade
glass facade

glass facade

Feedstock
electricity
-0.50

Feedstock fuels

-1.00

PV power retrofit systems integrated systems integrated systems


plant with heat recovery

Finally, the negative BOS contribution for PV integrated facades and roofs
should be remarked. This (theoretical) result reflects the possible use of PV
modules to replace conventional building cladding materials. The result is
7
particularly significant for the case of Alukobond panels . The energy needed to
manufacture a 1 mm thick aluminium foil is very high, larger than the BOS
energy content of a PV facade-integrated systems. As a consequence, the BOS
contribution is negative. The planning and design of a PV facade instead of an
Alukobond facade can be therefore considered as a conceptual energy saving
measure. Although purely theoretical, this result highlights the need for energy-
conscious architects and engineers to be aware of the hidden energy contents of
building materials.

All these results are even more significant in the case of future amorphous
silicon modules. In this case the EPBT of PV roof systems is always lower than

7. An Alukobond panel is made by a sandwich of two thin aluminium foils (total


thickness 1 to 3.5 mm) with a hard rubber layer in between. These panels are often
used in modern office buildings.

150
one year. With heat recovery it further drops down to less than six months. If the
substitution of Alukobond panels occurs, the total (theoretical) EPBT is zero!

2.5 Environmental benefits

Environmental benefits are evaluated here in terms of avoided emissions of


CO2. The indirect air emissions have been calculated according to the Italian
electricity production (0.531 kg CO2/kWhel) and distribution (0.567 kg CO2/kWhel)
8
mix [1]. For thermal energy production, a specific emission factor of 0.198 kg
CO2/kWhth has been taken into account (natural gas boilers).

As expected, at present the CO2 emissions produced during the manufacturing


and installation of all systems are significant, especially if compared with the
emissions avoided by the systems during their estimated life-time (25 years). This
can be also indicated in terms of CO2 yield ratio, defined as:
CO2 yield ratio = gross CO2 emissions avoided during lifetime
of PV system/CO2 emitted during production of PV system.

Today, conventional m-Si PV power plants save only 2.6 times the amount
of CO2 generated during their manufacturing, whereas for PV roofs with heat
recovery this value increases up to 5.4. It is worth recalling that results
concerning hybrid systems should be interpreted with some care, since they still
require more detailed investigations and further tests. More detailed LCAs of
hybrid systems are needed in the future, in order to take into account the
downstream use of the recovered heat.

In any case, the environmental benefits of PV systems in buildings will


significantly increase in future, as energy consumption and emissions during
manufacturing of modules will strongly decrease, and at the same time efficiencies
and lifetimes are expected to increase. Indeed, PV systems have a relevant
potential for improving their environmental performances. CO2 yield ratio values
range from present worst case of multi-crystalline silicon power plants (around 2)
to the best future cases of optimised multi-crystalline integrated silicon roofs (20
for simple roof; 34 for roofs with heat recovery; 38 if the substitution of
Alukobond panels is considered). Energy and CO2 yield ratios of future
amorphous silicon are in the range of 15-120 times (see also Table 2)

The figure also indicates that the environmental benefits of integrating PV


systems in buildings with respect to conventional PV power plants will

8. PV systems in buildings have no distribution losses, therefore their environmental


benefits are higher.

151
proportionally increase in the future. For example, the CO2 yield for an
integrated PV roof with heat recovery is expected to be around three times
higher than that of a conventional power plant.

The significant improvement achievable by PV systems can also be


expressed in terms of specific emissions during lifetime. Today, a
monocrystalline silicon PV power plant has a specific emission value of around
0.2 kg CO2/kWhel. This is mainly caused by indirect emissions deriving from
high (conventional mix) electricity consumption during manufacturing of
modules. In future, this value is expected to drop as low as 0.06 kg CO2 /kWhel
for PV power plants and 0.04 kg CO2/kWhel for integrated PV roofs [6].
Expected figures for future a-Si are even significantly lower.

2.6 External costs of PV

Within the ExternE project, two building-integrated systems have been


assessed in Germany. Table 4 summarises the main results.

As already mentioned in the earlier sections, a large share of the impacts


and related external costs is caused by airborne emissions of fossil fuel
electricity production. This reflects the fact that present crystalline PV
production technologies are not optimised and have a large consumption of
electricity. Of course much lower costs would result if one assumed that
electricity used for production was supplied by PV itself or other renewable
energy technologies. It is worth noticing that this is a feasible option nowadays
9
with the progressive liberalisation of energy markets.

In absolute terms, the damage estimates are low, but they are higher than the
ones for wind. Again, this result reflects the large consumption of electricity
produced by fossil fuels. Results based on rather high emission levels, reflecting the
present electricity production mix and the present PV technologies. Although not at
the highest level of estimated primary energy requirements shown in Figure 1, the
figures are rather high if one takes into account the rapid technological change of PV
production. This implies several recommendations for future calculations of external
costs of PV, which will be described in the next paragraph.

9. Environmental-oriented firms might want to purchase just green electricity

152
Table 3. Expected energy yield ratio of
different future optimised PV systems
2
Mean annual radiation: 1 700 kWhth/m on a 30° tilted, south-oriented surface;
lifetime: 30 years for crystalline silicon, 20 years for amorphous silicon

sc-Si mc-Si a-Si


PV plant 11.7 13.0 15.7
Retrofit flat roof 17.8 20.7 38.3
Retrofit tilted roof 17.2 20.1 42.9
Retrofit facade 11.8 13.9 23.4
Integrated tilted roof 20.4 24.3 71.3
Integrated glass facade 14.4 17.2 37.8
Integrated tilted roof with 33.3 39.7 123.5
heat recovery
Integrated glass facade 24.7 29.7 65.5
with heat recovery

Table 4. External costs of PV fuel cycle

MECU/kWh
Life cycle

Site, size Visual Global Human Other Sub-total


impact warning health
DE Emstal-Riede, ng 0.2-7.7 0.9 0.02 1.1-8.1 (1.9-3.3)
4.8 kW, Roof
DE Bielefeld, 13 ng 0.2-7.0 0.3 0.02 0.6-7.6 (1.4-2.8)
kW, Facade
(Source: ExternE Infosystem).

In the calculations carried out within ExternE, some benefits of building-


integration taken into account, i.e. the fact that building-integration implies no
land-use impacts. Furthermore it is assumed that good architectural integration
causes no visual impacts.

However, other advantages of building-integration are not taken into account.


For example, substitution of building cladding materials is not included. As we
have seen in the case of amorphous silicon facades, this can lead to (theoretical)
negative EPBT. Moreover, the multi-functional use of PV systems in buildings is
not considered. The possibility of heat recovery in mini-hybrid PV-Th collectors
for water pre-heating or air heating is not taken into account. Nor is the amount of

153
energy saving caused by sun-shading PV systems, decreasing the energy needs for
room cooling. All these factors might significantly reduce the external costs of PV,
in particular for future systems, for which the resource requirements and related
emissions are expected to decrease significantly.

On the other hand, potential impacts of other substances released during Si


purification (e.g. chlorosilanes) have been treated just qualitatively. This has to
be considered in future research, at least as long crystalline silicon PV industry
uses “classical”, (although optimised for larger quantities and lower purity)
silicon purification and feedstock preparation methods. However, it is also worth
noticing that several direct purification methods of metallurgical grade (MG-Si)
into solar-grade (SG-Si) silicon are currently experimented. This development
has to be taken into account very seriously, because the off-spec silicon scraps of
electronic industry are no longer sufficient to supply PV demand, due to the
rapid growth of PV worldwide market.

3. Wind

Wind is in some way similar to PV, in the sense that a significant share of
the environmental impacts is generated during the production phase. On the
contrary of PV however, other local impacts related to the power generation
phase, such as noise and visual amenity, are significant.

Within ExternE, 7 wind plants in 6 different countries have been assessed.


Table 5 summarises the main results.

As mentioned in the ExternE national implementation report, the damages


estimated are very low, the lowest of all fuel cycles studied. In fact, the higher
damages correspond to the indirect pollutant emissions produced during turbine
manufacturing. However, there are large variations among damages relate to
noise and visual impact, which means that wind plants have to be assessed
specifically for each site.

Apart from the LCI studies carried out within the ExternE project, several
other LCI or LCA studies of Wind systems have been published in past years
(see for instance [11,9]). All of them report the importance of the employed
materials for turbine manufacturing over the whole life-cycle.

154
Table 5. Overview of results of the wind fuel cycle

Site, size Power generation Other fuel cycle stages

Noise Visual Other Human Other Sub-total


impact health
DE Nordfriesland, 0.064 0.06 ng 0.31 0.03-1 0.37-1.3
11.25 MW (0.47-0.67)
DK Tunø Knob, 0.004 ng 0.009 0.5 0.1-3 0.6-3.6
5 MW off-shore (1-1.6)
DK Fjaldene, 0.02 0.2 0.02 0.3 0.1-2 0.6-2.5
9 MW (0.9-1.3)
ES Cabo Villano, 0.008 ng 0.95 0.8 0.02-0.7 1.7-2.7
3 MW (1.8-1.9)
GR Andros, 1.12 ng 0.14 0.9 0.03-1.14 2.2-3.3
1.6 MW (2.4-2.6)
NO Vikna, ng ng 0.003 0.4 0.06-2.1 0.5-2.5
2.2 MW (0.5-1.1)
UK Penrhyddlan, 0.07 ng 0.2 0.8 0.03-1.3 1.2-2.4
31 MW (1.3-1.5)
Source: ExternE, Vol. 10, “National Implementation”, p. 598.

In particular, I wish to mention and discuss the recent Environmental


Product Declaration (EPD) of a wind farm carried out by Sydkraft in Sweden.
On one hand an EPD has certainly to be acknowledged with great favour, for
various reasons, including:

• The commitment of a Private utility towards the use of LCA for external
communication and “green marketing” of its products and services.

• The fact that the multi-stakeholder approach of EPD, involving all


important actors from the very beginning of the certification process,
guarantees good rules on how to carry out the LCA and on how to report
its results. Credibility of results is further guaranteed via the certifying
institution.

However, with respect to the use of these data for the calculation of external
10
costs, there are also some limits:

• Results are reported in mid-points (e.g. kg of CO2-equivalents, kg of


CFC-11eq, etc.) and not in end-points.

10. Of course, this does not apply if the LCI data, which are available in principle only to
the certifier, are published entirely.

155
• In the Swedish EPD, results are reported in aggregated manner over the
whole life cycle. As far as this is concerned, it is worth mentioning that
the forthcoming pilot Italian EPD-system will require data to be
presented in clearly separate manner for each phase (production, use,
end-of-life).

Therefore, at the time-being, these data are of limited usefulness for the
calculation of external costs.

4. Future outlook

From the previous sections, it clearly appears that LCA is certainly


appropriate as a main supporting tool for the calculation of the external costs of
renewables. However, it is also clear that the current level of detail is not
enough, and that several refinements of analysis will be needed in the future.

4.1 PV

Present calculations of external costs of PV systems are clearly too limited.


In the future, many more studies shall be done. They should take into account
several aspects, including:

4.1.1 PV modules:
• Enlarging the scope of the analysis to other types of semiconductors, in
particular thin films.

• Taking into account the fast technological improvement of production


processes both for crystalline silicon cells and for thin film modules.

• Considering optimised silicon feedstock production methods, both based


on the “classical” Siemens process for the purification of MG-Si into
electronic-grade silicon (EG-Si) and on emerging alternative direct
purification processes of MG-Si into solar-grade silicon (SG-Si).

• As far the Siemens process is concerned, taking into account all impacts
of silicon purification, including the ones related to other emissions just
treated qualitatively in ExternE

• Including the dismantling and recycling phases of modules (some LCI


studies already exist).

156
4.1.2 Integration in buildings:

• Analysing different types of building integration and related BOS


efficiencies.

• Assessing the quality of data about employed materials for BOS.

• Making sensitivity analysis on the use of primary vs. recycled materials.

• Taking into account the potential multi-functional use of PV systems in


buildings (e.g. for sun-shading or heat recovery) and assessing the
relative credits.

4.1.3 Local aspects:

• Assessing local aspects (solar radiation and energy mix), which cannot
be generalised. Therefore, different cases shall be explicitly shown or
rules on how to make sensitivity analysis shall be provided.

• Making Sensitivity analysis on the use of the “conventional” energy mix


in a specific country vs. the use of hydro and/or wind for PV module
production. It is worth reminding that this a very concrete option within
the current and future market liberalisation process.

4.2 Wind

The crucial issue for LCA of wind systems is the quality of LCI data of
materials. This issue might be solved either by:

• The diffusion of more and more precise public or half-public11 data-


bases, such as the recent Italian National data-base I-LCA on materials,
energy, transport and waste management systems, or the IISI database
on steel production.

• The use of precise specific data, as foreseen for example in


Environmental Product Declarations (EPD).

Of course, local impacts of wind, e.g. noise and visual amenity, cannot be
generalised; therefore site-specific assessments shall be carried out.

11. Only for experts in the field.

157
5. The new project “eclipse”

To conclude I wish to briefly introduce a new research project in this area,


financed by the European Commission and called ECLIPSE (“ECological Life
cycle Inventories for present and future Power Systems in Europe”). ECLIPSE
will last 2 years, from January 2002 to December 2003 and will be carried out
by 8 research partners (5 research institutes and 3 utilities) from 7 different
European countries:

• Ambiente Italia - Istituto di Ricerche, Italy (co-ordinator).


• DLR - Deutsches Zentrum fuer Luft- und Raumfahrt, Germany.
• ESU Services, Switzerland.
• IER/Univ. Stuttgart (Germany).
• KEMA Nederland, The Netherlands.
• EdF (Électricité de France), France.
• Fortum, Finland.
• Vattenfall, Sweden.

The general objective of the research project ECLIPSE is to overcome


some current limitations of the use of Life Cycle Inventories (LCI) for
energy modelling and planning. More specifically the goal is to provide
potential users with:

• A coherent methodological framework, with application-dependent


methodological guidelines and data format requirements related to the
quantification of environmental impacts from power generation in
Europe based on a life cycle approach.

• A harmonised set of public, coherent, transparent and updated LCI data


on new and decentralised power systems, in a format which will make
them comparable to existing data of other energy technologies, easily
adaptable to local conditions and technological improvement and up-
datable.

In particular, the main objective is to fill the current gap of consistent and
reliable data on present and future new and decentralised technologies, by
creating a harmonised and coherent LCI data-set, which is of crucial importance
for any energy modelling, forecasting and planning. The work will cover about
100 different configurations of PV, wind, fuel cells, biomass and CHP

158
technologies. Sensitivity analysis to tackle with rapid technological
improvement and local conditions will be carried out. As far as this is concerned,
an accompanying manual illustrating the different technologies and showing
how to make sensitivity analysis will be provided.

Project results will increase the credibility, diffusion and exploitation of


LCI as a support tool for energy-environment-economy-modelling (like
ExternE) and planning as well as for other uses.

REFERENCES

[1] E.A. Alsema, Environmental Aspects of Solar Cell Modules, Summary


Report, Report 96074, Department of Science, Technology and Society,
Utrecht University, 1996.

[2] E.A. Alsema, Understanding Energy Pay-back Time: I Methods and


Results, IEA Expert Workshop on “Environmental Aspects of PV
Systems”, Utrecht, 1997.

[3] E.A. Alsema, B.C.W. van Engelenburg, Environmental Aspects and Risks
of Amorphous Silicon Solar Cells, Report, Dept. of Science, Technology
and Society of the University of Utrecht, The Netherlands, January 1994.

[4] E.A. Alsema, P. Frankl, K. Kato, Energy Pay-back Time of Photovoltaic


nd
Energy Systems : Present Status and Prospects, Proceedings of the 2 World
Photovoltaic Solar Energy Conference, Vienna, Austria, 6-10 July 1998,
p. 2125-2130, 1998.

[5] E.A. Alsema, E. Nieuwlaar, Energy Viability of Photovoltaic Systems,


Energy Policy, 28, p. 999-1010, 2000.

[6] P. Frankl, “Analisi del ciclo di vita di sistemi fotovoltaici, (Life-Cycle


Analysis of Photovoltaic Systems), Ph. D. dissertation thesis, Università di
Roma “La Sapienza”, Roma – available at the Biblioteca Nazionale, Roma
(Italy), or at the Dipartimento di Meccanica e Aeronautica, Università di
Roma “La Sapienza”, Roma (Italy), May 1996.

159
[7] P. Frankl, A. Masini, M. Gamberale, D. Toccaceli, Simplified Life-cycle
Analysis of PV Systems in Buildings : Present Situation and Future
Trends, Progress in Photovoltaics Res. Appl. 6, p. 137-146, 1998.

[8] Laboratory for Energy Systems (LES)/Swiss Federal Institute for


Technology, Zürich (ETH) and Paul Scherrer Institut (PSI),
Villigen/Wuerenlingen (Switzerland), Oekobilanzen fuer Energiesysteme
(Ecobalances of Energy Systems), R. Frischknecht, P. Hofstetter,
I. Knoepfel (ETH) & R. Dones, E. Zollinger (PSI) editors, Zürich
(Switzerland), 3rd edition, 1996.
[9] D.J. Gürzenich, N. Mathur, Kumar Bansal, H.-J. Wagner, Cumulative
Energy Demand for Selected Renewable Energy Technologies, The Int. J.
LCA 4 (3) 143-149, Ecomed publishers, 1999.

[10] G. Hagedorn, E. Hellriegel, Umwelrelevante Masseneintrage bei der


Herstellung verschiedener Solarzellentypen, Forschungstelle für
Energiewirtschaft, München (Germany), 1992.

[11] H. Holttinen, H. Mälkki, T. Turkulainen, H. Bijsterbosch, R. Schmidt, Life


Cycle Assessment of Different Wind Turbine Blade Materials, Proceedings
of EWEC’99, Nice, 1999.

[12] K.M. Hynes, A.E. Baumann, R. Hill, An assessment of Environmental


Impacts of Thin Film Cadmium Telluride Modules Based on Life Cycle
st
Analysis, l World Conf. on PV Energy Conversion, Hawaii, 1994.

[13] K.M. Hynes, N.M. Pearsall, M. Shaw, F.J. Crick, An Assessment of the
th
Energy Requirements of PV Cladding Systems, Proceedings of the 13
European Photovoltaic Solar Energy Conference, Nice (France),
23-27 October 1995.

[14] K. Kato, A. Murata, K. Sakuta, Energy Payback Time and Life-cycle CO2
Emission of Residential PV Power System with Silicon PV Module,
Progress in Photovoltaics, 6(2), p. 105-115, 1998.

[15] G.A. Keoleian, G.M. Lewis, Application of Life-cycle Energy Analysis to


Photovoltaic Module Design, Progress In Photovoltaics, 5, p. 287-300, 1997.

160
Session 3

COMPARATIVE ASSESSMENTS IN
ELECTRICITY AND TRANSPORTATION

161
LCA AND EXTERNAL COSTS IN
COMPARATIVE ASSESSMENT OF ELECTRICITY CHAINS.
DECISION SUPPORT FOR
SUSTAINABLE ELECTRICITY PROVISION?

Alfred Voss
Institute of Energy Economics
and the Rational Use of Energy, Stuttgart (Germany)

1. Introduction

The provision of energy and electricity plays an important role in a


country’s economic and environmental performance and the sustainability of its
development. Sustainable development of the energy and electricity sector
depends on finding ways of meeting energy service demands of the present
generation that are economically viable, environmentally sound, and socially
acceptable and do not jeopardize the ability of future generations to meet their
own energy needs.

As liberalised electricity markets are becoming widespread, according to


neo-classical welfare economics, getting the prices right is a prerequisite for
market mechanisms to work effectively towards sustainable development.

Life Cycle Assessment (LCA) and external cost valuation are considered to
offer opportunities to assist energy policy in a comprehensive comparative
evaluation of electricity supply options with regard to the different dimensions
of sustainable energy provision as well as in the implementation of appropriate
internalisation strategies.

The paper addresses life cycle assessment and external cost analysis carried
out for selected electricity systems of interest under German conditions. Results
from a comprehensive comparative assessment of various electricity supply
options with regard to their environmental impacts, health risks, raw materials
requirements as well as their resulting external cost will be summarised. The use
of LCA based indicators for assessing the relative sustainability of electricity
systems and the use of total (internal plus external) cost assessment as measure

163
of economic and environmental efficiency of energy systems will be discussed.
Open problems related to life cycle analysis of energy chains and the assessment
of environmental damage costs are critically reviewed, to illustrate how in spite
of existing uncertainties the state of the art results may provide helpful energy
policy decision support. The paper starts with some remarks on what the concept
of sustainability in terms of energy systems means.

2. The concept of sustainable development: What does it mean for the


energy system?

According to the Brundtland Commission, and the Rio Declarations, the


concept of “sustainable development” embraces two intuitively contradictory
demands, namely the sparing use of natural resources and further economic
development. The Brundtland Commission defines sustainable development as a
“development that meets the needs of the present generation without
compromising the ability of future generations to meet their own needs”.

Even if this definition has arisen against a background of environmental and


poverty problems, it nevertheless represents an ethically motivated claim which
is derived from considerations of fairness with future generations in mind. The
challenge is to simultaneously help to deliver economic prosperity, to reduce and
eventually eliminate poverty, to provide environmental quality and social equity
and to maintain the natural foundations of life.

Therefore, the aim of sustainable development is to bequeath future


generations with a stock of natural resources which will enable them to satisfy
their needs at least at the level we enjoy today. This general definition of
sustainability, which is acceptable to many, is not very specific about how we
can guarantee satisfying the needs of future generations, for example with
reference to the energy supply. It is both vague and open-ended and therefore
leaves room for different interpretations.

Any attempt to define the concept of sustainability in concrete terms can


only be sound if – as far as the material-energetic aspects are concerned – it
takes the laws of nature into account. In this context the second law of
thermodynamics which the chemist and philosopher Wilhelm Ostwald called
“The law of happening” [Das Gesetz des Geschehens] acquires particular
significance. The fundamental content of the second law of thermodynamics is
that life and the inherent need to satisfy requirements is vitally connected with
the consumption of workable energy and available material.

Within the context of defining the concept of sustainability in concrete


terms the need to limit ecological burdens and climate change can certainly be

164
substantiated. It becomes more difficult when confronted with the question of
whether the use of finite energy resources is compatible with the concept of
“sustainable development”, because oil and natural gas and even the nuclear
fuels which we consume today are not available for use by future generations.
This then permits the conclusion that only the use of “renewable energy” or
“renewable resources” is compatible with the concept of sustainability.

But this is not sound for two reasons. On the one hand the use of renewable
energy, e.g. of solar energy, also always goes hand in hand with a claim on non-
renewable resources, e.g. of non-energetic resources and materials which are
also in scarce supply. And, on the other hand, it would mean that non-renewable
resources may not be used at all – not even by future generations. Given that it
is, therefore, obviously impossible to pass on un-changed the non-renewable
resource base, the important thing within the meaning of the concept of
sustainable development is to bequeath to future generations a resource base
which is technically and economically usable and which allows their needs to be
satisfied at a level at least commensurate with that which we enjoy today.

However the energy and raw material base available is fundamentally


determined by the technology available. Deposits of energy and raw materials
which exist in the earth’s crust but which cannot be found or extracted in the
absence of the requisite exploration and extraction techniques or which cannot
be produced economically cannot make any contribution towards securing the
quality of life. It is therefore the state of the technology, which turns valueless
resources into available resources and plays a joint part in determining their
quantity. As far as the use of limited stocks of energy is concerned this means
that their use is compatible with the concept of sustainability as long as it is
possible to provide future generations with an equally large energy base which is
usable from a technical and economic viewpoint. Here we must note that in the
past the proven reserves, i.e. energy quantities available technically and
economically, have risen despite the increasing consumption of fossil fuels.
Moreover, technical and scientific progress has made new energy bases
technically and economically viable, for instance nuclear energy and part of the
renewable energy sources.

As far as the environmental dimension of sustainability is concerned, the


debate should take greater note of the fact that environmental pollution,
including those connected with today’s energy supply, are primarily caused by
anthropogenic flows of substances, by substance dispersion i.e. the release of
substances into the environment. It is not, therefore, the use of the working
potential of energy which pollutes the environment but the release of substances
connected with the respective energy system, for instance the sulphur dioxide or
carbon dioxide released after the combustion of coal, oil and gas. This becomes
clear in the case of solar energy which, with the working potential – solar

165
radiation – it makes available is, on the one hand, the principle source of all life
on earth but is also, on the other hand, by far the greatest generator of entropy,
because almost all of the sun’s energy is radiated back into space after it has
been devalued as heat at the ambient temperature. Since its energy, the radiation,
is not tied to a material energy carrier, the generation of entropy does not
produce any pollution in today’s sense of the word. This does not, of course,
exclude the release of substances and associated environmental pollution in
connection with the manufacture of the solar energy plant and its equipment.

The facts addressed here are of such particular significance because this
entails the possibility of uncoupling the consumption of energy and the pollution
of the environment. The increasing use of working potential (energy) and a
reduction in the burdens on the climate and the environment are not, therefore, a
contradiction in terms. It is the emission of substances that have to be limited, not
the energy flows themselves, if we want to protect the environment.

In addition to expanding the resource base available, the economical use of


energy or rather of all scarce resources is, of course, of particular significance in
connection with the concept of “sustainable development”. The efficient use of
resources in connection with the supply of energy does not only affect energy as
a resource since the provision of energy services also requires the use of other
scarce resources including, for instance, non-energetic raw materials, capital,
work and the environment. The efficient use of all resources as can be derived
from the concept of sustainability also corresponds to the general economic
principle, however. Both allow for the conclusion that an energy system or an
energy conversion chain for the provision of energy services is more efficient
than another if fewer resources, including the resource environment, are utilised
for the energy service.

In the economy costs and prices serve as the yardstick for measuring the use
on scarce resources. Lower costs with the same use mean an economically more
efficient solution which is more considerate on resources. The argument that can
be raised against using costs as a criterion for evaluating energy systems is that
the external effects of environmental damage for instance are not currently
incorporated in the cost-calculations. This circumstance can be remedied by an
internalisation of external costs. Without addressing the problems associated
with external cost valuation here, the concept of total social costs that is
combining the private costs with the external ones could serve as a suitable
yardstick for measuring the utilisation of scarce resources. Total social costs
could therefore serve as an integrated indicator of the relative sustainability of
the various energy and electricity supply options and it would be appropriate if,
in this function, they were again to be afforded greater significance in the energy
policy debate. Furthermore, cost efficiency is also the basis for a competitive
energy supply in order to secure the energy side of economic development and

166
adequate employment and it is also the key to avoiding intolerable climate
change. Both of these issues are central aspects of the concept of “sustainable
development”.

Following this clarification of the concept of sustainable development with


regard to the supply of energy we will now like to examine various electricity
production options as regards their contribution towards a sustainable
development of energy supply. The assessment will be based on a set of
sustainable development indicators, including emissions to the environment, the
requirement of both energetic and non-energetic non-renewable resources, health
impacts and economic performance.

3. LCA results – a first comparison of energy systems with a view to


sustainability

The approach of Life Cycle Assessment (LCA) provides a conceptual


framework for a detailed and comprehensive comparative evaluation of
electricity supply options with regard to their resource, health and environmental
impacts as important sustainability indicators. Full scope LCA considers not
only the direct emissions from power plant construction, operation and
decommissioning, but also the environmental burdens and resource requirements
associated with the entire lifetime of all relevant upstream and downstream
processes within the energy chain. This includes exploration, extraction, fuel
processing, transportation, waste treatment and storage. In addition, indirect
emissions originating from material manufacturing, the provision and use of
infrastructure and from energy inputs to all up- and downstream processes are
covered. As modern technologies increasingly tend to reduce the direct
environmental burdens of the energy conversion process, the detailed assessment
of all life cycle stages of the fuel chain is a prerequisite for a consistent
comparison of technologies with regard to sustainability criteria.

The LCA was carried out for a set of important electricity generation’s
option, which is considered as representative for near-future technologies to be
operated in Germany. Table 1 summarises some central technological parameters
for the selected reference technologies.

The following figures and tables will summarise results for some of the key
impact categories. Although based on our present level of knowledge this is not
a complete and comprehensive comparison of all the indicators that are
important from the point of view of sustainability, but it does provide an initial
indication of the potential contribution of specific electricity supply options to a
future sustainable energy system.

167
Table 1. Characterisation of the
reference electricity production technologies

Technology Power installed Efficiency Life

Coal Pulverised fuel firing 600 MW 43.0% 35 a

Lignite Pulverised fuel firing 800 MW 40.1% 35 a

Gas combined- Combined-cycle 777.5 MW 57.6% 35 a


cycle

Nuclear (PWR) Actual PWR 1375 MW 34.0% 40 a

PV (poly) Poly-crystalline 5 kW 9.5%1) 25 a


PV (amorphous) amorphous 5 kW 4.5%2) 25 a

Wind 5.5 m/s2) 1.5 MW – 20 a

Hydro Run-of-river 3.1 MW 90%3) 60 a


1) System-efficiency.
2) Average windspeed p.a.
3) Efficiency of turbines.

3.1 Cumulative energy requirements

The generation of electricity is associated with partly quite intensive energy


consumption by power plant construction, and – in the case of fossil and nuclear
energy sources – also by fuel supply and waste treatment. The cumulative energy
requirement as shown in Table 2 for different power generation systems includes
the primary energy demand for the construction and decommissioning of the
power plant as well as for the production and supply of the respective fuel. The
energy content of the fuel input is not included in the figures.

The indirect primary energy input per produced kWh of electricity for
hydro, wind and nuclear systems is in the range of 0.04 to 0.07 kWh. For natural
gas and coal the necessary energy input per produced unit of electricity is in the
range of 0.17 to 0.30 kWh which is basically determined by the energy required
for the extraction, transport and processing of the fuel. The corresponding
figures for today’s photovoltaic systems are 0.67 to 1.24 kWh. This is also
reflected in the energy amortisation time which is approximately 6 to 12 years in
the case of photovoltaic systems using today’s technology and is by far the
longest compared to any of the other systems.

168
Table 2. Cumulative energy requirements (CER)
and energy payback periods (EPP)

CER EPP
(without fuel) [months]
[kWhPrim/kWhel.]
Coal (43%) 0.3 3.6
Lignite (40%) 0.17 2.7
Gas CC (57.6%) 0.17 0.8
Nuclear (PWR) 0.07 2.9
PV (poly) 1.24 141
PV (amorph.) 0.67 76
Wind (5.5 m/s) 0.07 6.4
Hydro (3.1 MW) 0.04 10.9

3.2 Raw material requirements

Electricity production involves consumption of non-energetic raw materials


such as iron, copper or bauxite. Sustainability also means the efficient use of
such resources. Table 3 shows the cumulated resource requirements of the power
generation systems considered here for selected materials. It covers the raw
material requirements for power plant construction, fuel supply, and for the
supply of other raw materials. The table only includes a small part of the various
raw materials required and is therefore not a complete material balance.
However, results indicate that the relatively small energy density of solar
radiation and of the wind leads to a comparatively high material demand. This
high material intensity for wind and solar energy is an important aspect with
regard to the generation costs.

3.3 Pollutant emissions

Figure 1 compares the cumulative emissions of selected pollutants of the


power generation systems considered. It is obvious that electricity generated
from solid fossil fuels (hard coal and lignite) is characterised by the highest
emissions of SO2, CO2 and NOx per unit of electricity, while emissions from the
nuclear system, hydropower and wind are comparatively low. Electricity
generation from natural gas causes emissions that are significantly lower than
those from coal-fired systems.

169
Table 3. Total life cycle raw material requirements

Iron Copper Bauxite


[kg/GWhel.] [kg/GWhel.] [kg/GWhel.]
Coal (43%) 2 310 2 20
Lignite (40%) 2 100 8 19
Gas CC (57.6%) 1 207 3 28
Nuclear (PWR) 420-445 6 27
PV (poly) 5 350-7 300 240-330 2 040-2 750
PV (amorph.)
Wind (5.5 m/s) 3 700 50 32
Hydro (3.1 MW) 2 400 5 4

Although there are no direct emissions from the electricity generation stage,
the high material requirements for the production of PV panels result in
cumulative CO2 and NOx emissions of the photovoltaic fuel chain that are close
to those of the gas fuel chain and far higher as SO2 and particulates are
concerned.

It might be mentioned that the indirect emissions from material supply and
component manufacturing are determined to a great extent by the emissions of
the respective energy mix. Due to the high proportion of fossil energy in the
German electricity mix, results shown in Figure 1 are not directly applicable to
other countries with a different energy mix.

3.4 Human health risks

Electricity generation from fossil fuels, nuclear energy or renewable energy


sources leads to an increased level of air pollution, or to an increased exposure of
the population to ionising radiation, which in turn might cause an increased risk to
the health of the exposed population. Using the emissions from the life cycle
assessment as a starting point, health risks resulting from the operation of the
energy systems considered here are assessed following a detailed impact pathway
approach. For the quantification of health effects from pollutants relevant for fossil
energy systems (fine particles, SO2, Ozone) dose-effect models have been derived
from recent epidemiological literature. The risk factors recommended by the
International Commission on Radiological Protection (ICRP) are used to estimate
effects from ionising radiation.

170
Figure 1. Total life cycle emissions

130% 5
120%
110%
Life cycle emissions relative to lignite

2 2 2 2
100%
90%
1 1
1
80%
70%
60% 5 5
50% 3
40%
3
5 3 6
30% 1
4
20% 7 6 3 6
7 4
10% 6 7 4 7
4
0%
CO2 SO2 NOx PM10

1: Coal (min./max.) 2: Lignite (min./max.) 3: Gas CC (min./max.) 4: Nuclear (min./max.)


5: PV (min./max.) 6: Wind (min./max.) 7: Hydro (min./max.)

The application of the ICRP risk factors to the very small individual dose
resulting from long term and global exposure is, however, a matter of particular
uncertainty and might lead to an overestimation of effects. Results of the risk
assessment are summarised in the next figure. The increased death risk is
presented as the loss of life expectancy in Years of Life Lost (YOLL) per TWh.

Figure 2 shows that electricity generation from coal and lignite lead to the
highest health risks of the power generation systems considered, while power
generation from nuclear systems, wind and hydro energy is characterised by the
lowest risk. Due to the high emissions from the material supply, risks from
photovoltaic systems are higher than the risks from natural gas-fired power
plant. Results for the nuclear fuel chain include the expected value of risk from
beyond design nuclear accidents, which is small compared to the importance of
major nuclear accidents in the public discussion. However, the expected value of
risk is not necessarily the only parameter determining the acceptability of a
technology. Different evaluation schemes that take into account risk aversion or
a maximum tolerable impact might lead to a different ranking of technologies.

171
Figure 2. Health risks of energy systems

90
[YOLL/TWh] Total (max.)
80 Up- and downstream processes (min.)
Power plant emissions (min.)
70

60
Years of life lost

50

40

30

20

10

0 Coal Gas CC Nuclear PV Wind


Lignite Hydro

3.5 External costs

It is well accepted now that health impacts and environmental damage due
to air pollution cause economic losses which are not accounted for in the
electricity price (so called external costs). According to neo-classical welfare
economics, external costs have to be internalised, i.e. added to the price of
electricity, to achieve a full picture of the consumption of scarce resources.

External costs resulting from impacts on human health, agricultural crops


and building materials are considered as quantifiable with a reasonable level of
uncertainty, but impacts on ecosystems and in particular potential impacts from
global climate change are hardly quantifiable based on current knowledge, so
that an economic valuation of the potential impacts is very uncertain. In these
cases, marginal abatement costs for achieving policy-based environmental
targets (German CO2-reduction targets in the case of global warming, and SO2-
and NOx-targets derived from the European Commission’s strategy to combat
acidification for ecosystem protection) can be used to give a rough indication of
the potential damage costs. Using the detailed Life Cycle Inventories as guiding
input the marginal external cost estimates are based on applications of the
“impact pathway approach”, established in the EU ExternE Project. The “impact
pathway approach” models the causal relationships from the release of pollutants
through their interactions with the environment to a physical measure of impact
determined through damage functions and, where possible, a monetary valuation

172
of the resulting welfare losses. Based on the concept of welfare economies,
monetary valuation follows the approach of “willingness-to-pay” for improved
environmental quality. The valuation of increased mortality risks from air
pollution is based on the concept of “Value of Life Year Lost”.

External costs calculated for the reference technologies are summarised in


Figure 3. For the fossil electricity systems, human health effects, acidification of
ecosystems, and the potential global warming impacts are the major source of
external costs. Although, the power plants analysed are equipped with efficient
abatement technologies, the emission of SO2 and NOx due to the subsequent
formation of sulphate and nitrate aerosols leads to considerable health effects
due to increased “chronic” mortality. A comparison between the fossil systems
shows that health and environmental impacts from the natural gas combined
cycle plant are much lower than from the coal and the lignite plant.

Figure 3. External costs from different


electricity generation technologies operated in Germany
4

3.5

3
Euro-Cent/kWh

2.5

1.5

0.5

0
ph
te

ly

ro
r

d
l

a
oa

or

po

in
i

yd
le
C
gn
C

W
am
uc
as

H
PV
Li

N
G

PV

Acidification/Eutrofication1) Global warming2)


Health effects Material damage & Others3)

1) Acidification/eutrofication: valuation based on marginal abatement costs required to


achieve the EU “50%-Gap Closure” target to reduce acidification in Europe.
2) Global warming: valuation based on marginal CO2-emissions in Germany by 25% in
2010 (19 Euro/tCO2).
3) Others include noise and crop losses.

External costs arising from the nuclear fuel chain are significantly lower
than those estimated for the fossil fuels. Most of the radiological impacts are
calculated by integrating very small individual doses over 10 000 years. The
application of the ICRP risk factors in this context is at least questionable, and
most likely leads to an overestimation of effects. The impact resulting from

173
emissions of “conventional” (i.e. SO2. NOx, and particles) air pollutants from the
nuclear fuel chain dominate the external costs. The external costs calculated
from the expected value of risk from beyond design nuclear accidents are
surprisingly small compared to the importance of major nuclear accidents in the
public discussion.

External cost of photovoltaic, wind and hydropower mainly result from the
use of fossil fuels for material supply and during the construction phase.
External costs from current PV application in Germany are higher than those
from the nuclear fuel chain and close to those from the gas fired power plant.
Impacts from the full wind and hydropower life cycle are lower than those from
all other systems, thus leading to the lowest external costs of all the reference
technologies considered. While the uncertainties in the quantification of external
costs are still relatively large, the ranking of the considered electricity options is
quite robust.

3.6 Power generation costs

Costs in general might be considered as a helpful indicator for measuring the


use of sparse resources. It is thus not surprising that a high raw material and
energy intensity is reflected in high costs. The power generation costs shown in
the next figure indicate that power generation from renewable energies is
associated with higher costs – much higher in the case of solar energy – than those
resulting from fossil-fired or nuclear power plants. However, as discussed above,
the private costs alone do not fully reflect the use of scarce resources. To account
for environmental externalities, external costs have to be internalised, i.e. added to
the private generation costs. Figure 4 shows that the external costs resulting from
the electricity generation of fossil fuels amount from 30% (natural gas) to about
100% (lignite) of the generation costs, while for the other technologies the external
costs are only a small proportion of generation costs.

Figure 4 shows that the external costs resulting from the electricity generation
of fossil fuels amount from 30% (natural gas) to about 100% (lignite) of the
generation costs, while for the other technologies the external costs are only a
small proportion of generation costs. The internalisation of external costs might
lead to competitiveness of some wind and hydropower sites compared to fossil
fuels, but do not affect the cost ratios between the renewable and the nuclear
systems. On the other hand it is obvious, that the full internalisation of
environmental externalities would improve the competitive advantage of nuclear
energy to fossil electricity production.

The results of energy and raw material requirements, life cycle emissions,
risks and both external and generation costs discussed so far are based on the

174
characteristics of current technologies. It is expected that technical development
will result in a further reduction in costs and in the environmental burdens of
power generation. However, this applies to all the power generation technologies
considered here. Preliminary results for future systems indicate that the ranking
of technologies with respect to total costs is quite robust.

Figure 4. Total costs of various electricity


generation technologies operated in Germany

Photovoltaics
(polycrystalline

Hydroelectric power

Wind

Hardcoal1)
Min. Max.
1) Production costs
Lignite

Natural gas1) External costs

Nuclear power1)

0 5 10 15 20 50 55 60 65 70 75 80
Costs of electricity production and external costs [Euro-Cent/kWel]
1) Base-load.

4. Uncertainties and open problems

In spite of considerable progress that has been made over the last years,
especially with regard to the bottom up modelling of the full impact pathways
and the monetary valuation of health effects, the life cycle inventory based
quantification and valuation of environmental impacts is still linked to partly
large uncertainties. Uncertainties related to impact assessment and valuation are
certainly larger compared to those of the LCA inventory. Besides the data and
model uncertainties the estimation of external cost is faced with systematic
uncertainties arising from lack of knowledge, and value choices that influence
the results.

Lack of knowledge is the single most important reason for the large
uncertainties related to the quantification of climate change damage costs. This
suggests using abatement costs based on the standard-price approach to achieve
a specific greenhouse gas reduction target as the second-best solution to make

175
this impact visible in the external cost estimates. Large uncertainties are in the
exposure-response functions for health impacts and the “Value of Life Year
Lost”. As in the past improved knowledge, e.g. of the influence of particle
composition on the chronic mortality from fine particles could lead to different
health related damage costs. But an assessment can always only reflect the
current knowledge.

The estimates of external costs are influenced also by the discount rate
chosen, to account for damage costs in the future as well as by valuation of
damages in different parts of the world. The uncertainties stemming from these
assumptions can be best dealt with by sensitivity analysis.

The application of the impact pathway approach and the monetary valuation
methods suggest relatively low external costs for both beyond design accidents
in a nuclear power plant and radioactive waste deposit. The impact from a single
beyond design accident can be very large, but normalised to the electricity
generation over the power plants lifetime, the expected value of risk (i.e.
probability times consequences) is low, a fact which is even robust against
uncertainties in the accident probability.

Some people argue that the use of the expected value of risk to estimate the
external cost of a low probability events with large consequences is an open
problem. They consider the maximum damage from a single incident as an
important key criterion on its own, which has to be included in the impact
valuation of technologies. Empirical evidence supporting this kind of reasoning
is still missing.

With respect to the uncertainties of external cost estimates it is important to


notice, that these uncertainties are of relevance to any other valuation scheme. It
is however remarkable, that in spite of these uncertainties and changing
background assumptions, external cost estimates at least indicate a robust
relative ranking for the key electricity production technologies. However, care
has to be taken to acknowledge existing uncertainties and to not take external
cost estimates out of the given context when using them in a policy context.

5. LCA and external costs for policy support

Life Cycle Assessment and the external as well as the total cost approach
can provide valuable decision support for a wide range of policy relevant issues:

• Assessment of technologies currently used to identify deficiencies and


potentials for improvement and corresponding research issues.

176
• Comparison of current and future energy supply options with respect to
their health and environmental impacts, resource requirements and with
respect to their compliance with sustainability indicators.

• Cost-benefit-analysis of environmental policy measures.

• Extension of national green-accounting frameworks.

There are several examples of successful applications in these areas, for


instance the use of the method for cost-benefit analysis of desulphurisation plants
attached to large coal fired power plants in Europe. The comparison of the costs in
per tonne of SO2 avoided and the avoided damage costs due to the reduced
emissions shows that the benefit clearly outweighs the costs. This even remains
valid, if the mortality impacts – the impact category considered to involve the
highest uncertainty – would be neglected. So, even if the damage cost estimates
are varied within the uncertainty range, the conclusion would not change.

In the context of the liberalisation of the European electricity market the


concept of internalising external effects by means of technology-specific price
adders has been discussed. The idea is to derive science based recommendations
on the height of price adders for electricity production by different technologies
from LCA and external cost research. This approach is expected to get the prices
right in competitive markets and to ensure that the use of the environment is
accounted for in the market mechanism.

But the use of simple price adders for each technology appears inappropriate,
besides the still existing large uncertainties of the external cost estimates. This is
due to the fact that health and environmental impacts depend heavily on the
concrete technical design and the location of a specific power plant. For this
reason the use of regionally differentiated pollutant-specific damage costs is
recommended for the internalisation of external costs due to airborne pollutants.
These damage costs should at least be differentiated by country. Table 4 presents
damage costs per tonne of pollutant emitted in Germany.

Table 4. Specific damage costs in per tonne of pollutant


emitted in Germany (reference year 1998)

per tonne emitted


SO2 5 650
NOx 5 030
PM10 8 700
NMVOC 1 770

177
The advantage of such a pollutant-oriented approach is that it gives a direct
incentive for reducing the emissions. This is expected to outweigh the
disadvantage of a higher effort for recording the emissions of every pollutant.

In the case of CO2 and other greenhouse gases, due to the high uncertainties
involved in estimating damage costs, it is recommended to base internalisation
instruments (such as emission certificates or a greenhouse gas tax) on greenhouse
gas reduction targets and the associated marginal abatement cost.The risks of low
probability accidents should be integrated into the monetary accounting system
by introducing liability insurance obligations.

It can be concluded that LCA and quantification of environmental costs give


valuable input to the assessment of the relative sustainability of different electricity
production technologies in spite of the current knowledge gaps. These methods thus
can contribute to a rational decision support. LCA inventories provide very useful
information concerning the evaluation of resource requirements of different
electricity production options. External cost estimates represent an aggregated
indicator of environmental performance. Together with the private costs, total costs
can serve as an integrated indicator for the overall resource consumption and in this
respect for relative sustainability of the different energy options.

As far as applicable, the approach of monetary valuation can be considered to


be the most appropriate way of weighting and aggregating different impact
categories when assessing the environmental impact or the resource intensity of
energy systems. Monetary values based on individual and social preferences for a
wide range of health and environmental impacts have been derived from empirical
work. The use of such values for aggregation has advantages compared to weighting
schemes derived from expert or personal judgement, as the weighting is based on
“measured” preferences. And last, but not least, monetary values have the advantage
of being more illustrative than “utility points” or other artificial measures, although
the results might be the same.

178
BIBLIOGRAPHY

P. Bickel, S. Schmid, W. Krewitt, R. Friedrich, (1999), Monetary Valuation of


Life Cycle Impacts: the Case of Road and Rail Transport in Germany, Paper
presented at the 9th Annual Meeting of SETAC-Europe, 25-29 May 1999,
Leipzig (Germany), 1999.

European Commission, (1995a), ExternE – Externalities of Energy, Vol. 1:


Summary (EUR 16520), Luxembourg, Office for Official Publications of the
European Communities, 1995.

European Commission, (1995b), ExternE – Externalities of Energy, Vol. 2:


Methodology (EUR 16521), Office for Official Publications of the European
Communities, Luxembourg, 1995.

European Commission, (1995c), ExternE – Externalities of Energy, Vol. 3, Coal


and Lignite (EUR 16522), Office for Official Publications of the European
Communities, Luxembourg, 1995.

European Commission, (1995d), ExternE – Externalities of Energy, Vol. 4: Oil


and Gas (EUR 16523), Office for Official Publications of the European
Communities, Luxembourg, 1995.

European Commission, (1995e), ExternE – Externalities of Energy, Vol. 5:


Nuclear (EUR 16524), Office for Official Publications of the European
Communities, Luxembourg, 1995.

European Commission, (1995f), ExternE – Externalities of Energy, Vol. 6: Wind


and Hydro Fuel Cycles (EUR 16525), Office for Official Publications of the
European Communities, Luxembourg, 1995.

European Commission, (1999a), ExternE – Externalities of Energy, Vol. 7,


Methodology 1998 Update (EUR 19083), Office for Official Publications of the
European Communities, Luxembourg, 1999.

179
European Commission, (1999b), ExternE – Externalities of Energy, Vol. 8: Global
Warming (EUR 18836), Office for Official Publications of the European
Communities, Luxembourg, 1999.

European Commission, (1999c), ExternE – Externalities of Energy, Vol. 9, Fuel


Cycles for Emerging and End-use Technologies, Transport and Waste (EUR
18887), Luxembourg, Office for Official Publications of the European
Communities, 1999.

European Commission, (1999d), ExternE – Externalities of Energy, Vol. 10,


National Implementation (EUR 18528), Luxembourg, Office for Official
Publications of the European Communities, 1999.

R. Friedrich, P. Bickel, (Eds.), (2001), Environmental External Costs of


Transport, Springer-Verlag Berlin, Heidelberg, New York, 2001.

R. Friedrich, W. Krewitt, (1998), Externe Kosten der Stromerzeugung, in:


Energiewirtschaftliche Tagesfragen, 48. Jg. Heft 12, S. 789-794, 1998.

V. Hauff. (Hrsg., 1987), Unsere gemeinsame Zukunft: Der Brundtland-Bericht


der Weltkommission für Umwelt und Entwicklung. Greven, 1987.

S. Hirschberg, R. Dones, (2000), Analytical Decision Support for Sustainable


Electricity Supply, In: Proceedings of VDI Conference on Energy and
Sustainable Development: Contributions to Future Energy Supply, VDI,
Düsseldorf, Germany, pp. 168-187, 2000.

K. Knizia, (1992), Kreativität, Energie und Entropie: Gedanken gegen den


Zeitgeist, ECON Verlag, 1992.

W. Krewitt, R. Friedrich, T. Heck, P. Mayerhofer, (1998), Assessment of


Environmental and Health Benefits From the Implementation of the UN-ECE
Protocols on Long Range Transboundary Air Pollution. Journal of Hazardous
Materials 61, 239-247, 1998.

W. Krewitt, M. Holland, A. Trukenmüller, T. Heck, R. Friedrich, (1999), Comparing


Costs and Environmental Benefits of Strategies to Combat Acidification in Europe.
Environmental Economics and Policy Studies, (2) 4, 1999.

W. Krewitt, F. Hurley, A. Trukenmüller, R. Friedrich, (1998), Health Risks of


Energy Systems. Int. Journal of Risk Analysis, 18, No. 4, 1998.

180
W. Krewitt, P. Mayerhofer, A. Trukenmüller, R. Friedrich, (1998), Application
of the Impact Pathway Analysis in the Context of LCA – The Long Way from
Burden to Impact. In: International Journal of Life Cycle Assessment 3 (2),
pp. 86-94, 1998.

W. Krewitt, A. Trukenmüller, T.M. Bachmann, T. Heck, (2001), Country-specific


Damage Factors for Air Pollutants: A Step Towards Site Dependent Life Cycle
Impact Assessment. In: The International Journal of Life Cycle Assessment 6 (4),
pp. 199-210, 2001.

T. Marheineke et al., (2000), Ganzheitliche Bilanzierung der Energie- und


Stoffströme von Energieversorgungstechniken, IER-Forschungsbericht Band 74,
Institut für Energiewirtschaft und rationelle Energieanwendung, Stuttgart, 2000.

T. Marheineke, R. Friedrich, W. Krewitt, (1999), Application of a Hybrid-approach


to the Life Cycle Inventory Analysis of Freight Transport Tasks, SAE 1998
Transactions – Journal of Passenger Cars. Society of Automotive Engineers (SAE),
Warrendale PA (USA), 1999.

A. Voß, (1997), Leitbild und Wege einer Umwelt- und Klimaverträglichen


Energieversorgung. In: H.G. Brauch (Hrsg.) Energiepolitik, Springer Verlag, 1997.

A. Voß, A. Greßmann, (1998), Leitbild “Nachhaltige Entwicklung”– Bedeutung


für die Energieversorgung. In: Energiewirtschaftliche Tagesfragen, 48. Jg., Heft 8,
S. 486-491, 1998.

A. Voß, (1999), Sustainability in the Energy Supply-relevant Power Generation


Processes on the Test Rig, VGB Power Tech , No. 1, pp. 21-25, 1999.

181
LIFE-CYCLE ANALYSIS AND EXTERNAL COSTS
IN TRANSPORTATION

Mark A. Delucchi
University of California, California, USA

Overview
• A life cycle analysis of emissions of
air pollutants and greenhouse gases
from transportation fuels.
• External costs of motor-vehicle use

• Social cost comparison of alternative


vehicles and transportation modes

183
LCA References
• M.A. Delucchi, “A Lifecycle Emissions Analysis: Urban
Air Pollutants and Greenhouse-Gases from Petroleum,
Natural Gas, LPG, and Other Fuels for Highway Vehicles,
Forklifts, and Household Heating in The U S”, World
Resources Review, in press (2001).
• M.A. Delucchi, “Transportation and Global Climate”,
Journal of Urban Technology 6 (1): 25-46 (1999).
• M.A. DeLucchi, “Emissions from the Production, Storage,
and Transport of Crude Oil and Gasoline”, Journal of the
Air and Waste Management Association 43: 1486-1495
(1993).

External Cost References


• M. A. Delucchi, “Environmental Externalities of Motor-Vehicle Use in the
US”, Journal of Transport Economics and Policy 34: 135-168, May (2000).

• J.J. Murphy and M.A. Delucchi, “A Review of the literature on the Social
Cost of Motor-Vehicle Use in the United States”, Journal of Transportation
and Statistics 1 (1): 15-42, January (1998).

• M.A. Delucchi, “The Social Cost of Motor-Vehicle Use”, The Annals of the
American Academy of Political and Social Science 553: 130-142 (1997).

• M.A. Delucchi, “Should We Try to Get the Prices Right?”, Access, Number
16, University of California Transportation Center, Berkeley, pp. 14-21,
Spring (2000).

184
An LCA of Transportation Fuels
Fuels and electricity life cycle Vehicles and infrastructure life cycle
• End use of fuel • Materials production
• Dispensing of fuels • Vehicle assembly
• Fuel distribution • Operation and maintenance
• Secondary fuel cycle for transport
• Fuel production
modes
• Feedstock transport • Infrastructure construction
• Feedstock production

Why is LCA important?


Compare CO2 emissions from end use vs fuel cycle, for motor
vehicles (as a percentage of fossil-fuel CO2):

End use Whole fuel-cycle


US 22 30
OECD-Europe 18 24
World 14 19

185
Upstream emissions from alternative transportation fuel
cycles in year 2010 (g/106-BTU-fuel)

Fuel--> EtOH LPG Diesel CNG RFG MeOH CH2 EtOH


Feed--> wood oil&NG oil NG oil NG NG corn
CO2 (9,470) 8,688 14,672 10,167 20,842 29,788 90,662 70,174
NMOCs 32.3 16.2 14.9 6.5 43.9 25.4 9.6 230.7
CH4 72.0 148.5 209.3 285.9 220.7 342.2 379.2 198.7
CO 185.8 37.8 60.2 33.8 58.6 68.1 58.4 199.1
N2O 19.8 0.3 0.8 0.3 0.6 1.1 1.5 52.3
NOx 258.6 51.7 71.3 63.1 80.2 153.7 132.5 412.4
SOx 10.9 27.1 54.7 15.6 50.4 38.0 52.4 94.9
PM 43.5 8.7 18.6 3.6 18.3 8.2 9.3 122.4

CO2eq (1,095) 11,436 18,410 15,787 24,876 36,490 97,953 91,467

Upstream emissions from alternative transportation


fuelcycles in year 2010 (% of end-use emissions)

Fuel--> EtOH LPG Diesel CNG RFG MeOH CH2 EtOH


Feed--> wood oil&NG oil NG oil NG NG corn
CO2 -15% 14% 21% 20% 31% 48% 7430% 108%
NMOCs 26% 39% 21% 42% 28% 34% 77% 187%
CH4 392% 1270% 4596% 164% 1958% 5592% 6663% 1082%
CO 16% 3% 8% 3% 4% 6% 16% 17%
N 2O 53% 1% 25% 1% 2% 3% 140%
NOx 129% 27% 8% 33% 46% 77% 63% 206%
SOx 180% 636% 276% 560% 777% 612% 894% 1569%
PM 1426% 475% 51% 249% 261% 269% 547% 4010%

CO2eq -1% 15% 25% 23% 29% 46% 4151% 109%

186
Emissions from the vehicle life cycle,
year 2010
(% of end
Pollutant Emissions (g/lb) Emissions (g/mi) Emissions
use)

LDGVs HDDVs LDGV HDDV LDGVs HDDVs

CO2 3,838 3,285 83 161 30% 5%


NMOCs 1.91 1.85 0.04 0.09 5% 3%
CH4 8.03 6.93 0.17 0.34 381% 177%
CO 7.19 8.14 0.15 0.40 2% 1%
N2O 0.09 0.08 0.00 0.00 1% 3%
NOx 9.51 8.70 0.20 0.43 25% 1%
SOx 10.18 10.05 0.22 0.49 280% 59%
PM 6.99 7.11 0.15 0.35 532% 23%
CO2eq 3,865 3,293 83 161 23% 5%

Lifecycle GHG emissions from light-duty vehicles


(g/mi and % change)
Fuel cycle only Fuel+materials

Baseline gasoline vehicle 455.1 554.4

Diesel (0.032% S) -34.6% -30.5%


Natural gas (CNG) -27.5% -21.3%
LPG (P95/BU5) -23.4% -19.2%
Ethanol from corn -10.6% -8.7%

Ethanol wood and grass -76.5% -62.7%

Battery EV, coal plants -16.5% -6.2%

Battery EV, NG plants -60.4% -42.2%

FCEV, Methanol from NG -46.7% -38.7%

FCEV, Hydrogen from water -92.1% -75.9%

FCEV, Hydrogen from NG -61.4% -50.8%

187
External cost of motor-vehicle emissions in urban areas of the
US (10% change in emissions) (1991 $/kg)

PM10 NOx SOx CO VOCs, O3


Low High Low High Low High Low High Low High
Health 1.5 187.5 1.5 23.3 4.4 90.9 0.0 0.1 0.1 1.6
Visibility 0.1 5.2 0.2 1.5 0.5 5.3 0.0 0.0 0.0 0.1
Crops n.e. n.e. 0.0 0.0 n.e. n.e. 0.0 0.0 0.2 0.3
Forests, materials n.e. n.e. 0.0 0.0 n.e. n.e. 0.0 0.0 0.2 0.3
Climate change 0.0 0.000 1.20 20.0 0.0 0.000 0.60 10.0 1.5 25.0
($/Mg)
Total MVs 14.3 192.7 1.9 24.9 10.9 96.3 0.0 0.1 0.5 2.3
Total MVs+U 12.8 162.9 1.8 23.5 4.9 37.1 0.0 0.1 0.5 2.0
Total 1.6 32.8 1.8 23.5 4.9 37.1 0.0 0.1 0.5 2.0
MVs+U+RD

External cost of oil use in the US, 1991 $/end-use gallon


Gasoline Diesel All
vehicles vehicles vehicles
Strategic Petroleum Reserve - low 0.0004 0.0006 0.0005
Strategic Petroleum Reserve - high 0.0052 0.0064 0.0054
Defense costs - low 0.0045 0.0056 0.0047
Defense costs - high 0.0505 0.0623 0.0529
Pecuniary externality - low 0.0285 0.0350 0.0298
Pecuniary externality - high 0.0596 0.0730 0.0623
Price-shock cost to GNP - low 0.0189 0.0231 0.0198
Price-shock cost to GNP - high 0.1889 0.2314 0.1976
Water pollution - low 0.0023 0.0026 0.0023
Water pollution - high 0.0076 0.0084 0.0078
All costs - low 0.055 0.067 0.057
All costs - high 0.312 0.382 0.326

188
The marginal cost of noise from a 10% increase in VMT, for different types
of vehicles on different types of roads, in urbanized areas (1991$1000-VMT)
B ase I n te r s t a t e O th e r P r in c i p a l M in o r C o l l e c to r s Local
case free w a ys a r te r i a ls a r t e r i a ls ro ad s
LD A s 2 .9 6 4 .2 5 1 .1 8 0 .5 7 0 .0 7 0 .0 0
M D Ts 8 .5 0 1 3 .2 0 7 .0 2 5 .3 7 1 .0 5 0 .0 0
HDTs 1 6 .6 9 3 0 .8 0 2 0 .0 7 2 9 .9 3 4 .9 3 0 .0 0
B u ses 6 .3 6 9 .7 7 7 .1 8 6 .4 2 1 .2 2 0 .0 0
MCs 1 7 .1 5 2 7 .0 3 8 .7 1 4 .6 7 0 .5 6 0 .0 0
L o w case
LD A s 0 .1 1 0 .1 8 0 .0 4 0 .0 1 0 .0 0 0 .0 0
M D Ts 0 .4 0 0 .6 6 0 .3 2 0 .1 8 0 .0 1 0 .0 0
HDTs 0 .8 1 1 .6 2 1 .2 2 1 .7 7 0 .0 6 0 .0 0
B u ses 0 .3 5 0 .5 8 0 .3 8 0 .2 2 0 .0 0 0 .0 0
MCs 0 .6 6 1 .1 3 0 .2 7 0 .0 9 0 .0 0 0 .0 0
H ig h ca se
LD A s 4 0 .1 1 5 6 .0 2 1 6 .2 0 9 .3 5 6 .0 4 0 .4 4
M D Ts 1 1 4 .7 6 1 7 3 .3 8 9 6 .0 5 8 4 .9 3 7 8 .8 4 1 2 .1 3
HDTs 2 2 5 .6 1 4 0 4 .8 2 2 6 9 .2 7 4 1 4 .1 7 3 1 9 .2 2 9 2 .0 4
B u ses 8 6 .1 5 1 2 8 .6 0 9 8 .6 6 1 0 5 .3 3 1 0 8 .0 0 1 2 .8 4
MCs 2 3 2 .4 7 3 5 5 .7 3 1 1 9 .6 4 7 6 .6 5 5 0 .0 8 2 .7 3

Non-monetary externalities of motor-vehicle use in the US


(109 1991 $)

Cost item Low High Q


Accidental pain, suffering, and death, not accounted for by economically responsible party 10.2 120.0 A3, D
Travel delay, imposed by other drivers (includng accidents) that displaces unpaid activities 30.8 119.5 A2

Air pollution: human mortality and morbidity due to particulate emissions from vehiclesb 16.7 266.4 A1

Air pollution: human mortality and morbidity due to all other pollutants from vehicles 2.3 17.1 A1
Air pollution: human mortality and morbidity, due to all pollutants from upstream processes 2.3 13.0 A1
Air pollution: human mortality and morbidity, due to road dust 3.0 153.5 A1
Air pollution: loss of visibility, due to all pollutants attributable to motor vehicles 5.1 36.9 A1
Air pollution: damage to agricultural crops, due to ozone attributable to motor vehicles 2.1 3.9 A1
Air pollution: damages to materials, due to all pollutants attributable to motor vehicles 0.4 8.0 B [A1]
Air pollution: damage to forests, due to all pollutants attributable to motor vehicles 0.2 2.0 B [A2]
Global warming due to fuel-cycle emissions of greenhouse gases (U. S. damages only) 0.7 7.4 A1, B
Noise from motor vehicles 0.5 15.0 A1
Water pollution: health and environmental effects of leaking motor-fuel storage tanks 0.1 0.5 D
Water pollution: environmental and economic impacts of large oil spills 2.0 5.0 C [A1]
Water pollution: urban runoff polluted by motor-vehicle oil, and by highway de-icing 0.7 1.7 D
Pain and suffering and other non-monetary costs due to crimes related to motor-vehicle use 1.7 6.1 A3
Nonmonetary costs of injuries and deaths caused by fires related to motor-vehicle use 0.0 0.2 A3

189
External costs of EVs versus gasoline
vehicles (cents/mile)

Battery EVs Gasoline ICEVs


low high best low high best
Noise 0.00 1.20 0.04 0.00 1.60 0.05
Externalities of oil use 0.02 0.12 0.04 0.22 1.25 0.40
Climate change 0.00 0.13 0.06 0.01 0.19 0.09
Air pollution 0.02 0.21 0.07 0.19 2.32 0.75
TOTAL 0.05 1.67 0.21 0.41 5.36 1.29

Social cost of EVs vs. gasoline vehicles


(cents/mi)
Difference in costs
low high best
Private lifecycle costs 0.0 30.00 10.00
Noise 0.00 -0.40 -0.01
Externalities of oil use -0.20 -1.12 -0.36
Climate change -0.00 -0.06 -0.03
Air pollution -0.17 -2.11 -0.69
Total externalities -0.37 -3.69 -1.09
Social cost -4 30 9

190
External costs and subsidies for different passenger
transport modes
(cents per vehicle mile, except last row is cents per passenger mile)
[Numbers in brackets are my best estimates]

Cost item Gas auto Electric auto Transit bus Light rail Heavy rail
Air pollution [2.0] 0.8 to 13 1.5 [20.0] 5.4 to 123 5? 5?
Oil use, water pollution [0.8] 0.3 to 1.5 0.4 [4.0] 1.5 to 8.7 1? 1?
Noise [0.2] 0.01 to 2.0 0.15 [2.0] 0.5 to 10.0 1? 1?
Congestion 4.0 4.0 8.0 n.e. n.e.
Accidents 2.5 2.5 3.5 2? 2?
Marginal highway and service costs 0.1 0.1 1.5 0 0
Unpriced parking [1?] 0 to 8 [1?] 0 to 8 0 0 0
Inefficient highway user taxes and -2.7 0 0 (exempt from 0 0
fees, meant to cover highway costs fuel taxes)
Government subsidy: operating costs 0 0 339, 398, 465 685, 1137, 2800 372, 797, 1177
minus fares, operating+rolling-stock [398]
costs minus fares, total
operating+capital costs minus fares
Extra private costs relative to gas auto 0 0 to 16 [8] see subsidy see subsidy see subsidy
Total cents per vehicle-mile [8] 5 to 28 [18] 9 to 25 359 to 620 [437] 694 to 2,809 381 to 1,186
Passengers per vehicle assume 1.0 assume 1.0 11 (average) 26 (average) 22 (average)
Total cents per passenger-mile [8] 5 to 28 [18] 9 to 25 33 to 57 [40] 27 to 108 17 to 53

Conclusions
• LCA is important in transportation, because
“upstream” impacts can be significant
• Environmental external costs are dominated by the
health costs of particulate air pollution
• In the comparison of the social cost of
transportation alternatives, differences in external
cost are not trivial, but often are small compared
with differences in private costs or in financial
subsidies

191
Round Table

HOW TO USE INTERNALISATION OF


EXTERNALITIES IN POLICY MAKING?

193
ENERGY POLICY AND EXTERNALITIES:
THE LIFE CYCLE ANALYSIS APPROACH

1
Maria Rosa Virdis
International Energy Agency, Paris (France)

1. Introduction

In the energy sector, getting the prices right is a prerequisite for market
mechanisms to work effectively towards sustainable development. However,
energy production and use creates “costs” external to traditional accounting
practices, such as damages to human health and the environment resulting from
residual emissions or risks associated with dependence on foreign suppliers.
Energy market prices do not fully reflect those external costs. For example, the
costs of climate change are not internalised and, therefore, consumers do not get
the right price signals leading them to make choices that are optimised from a
societal viewpoint.

Economic theory has developed approaches to assessing and internalising


external costs that can be applied to the energy sector and, in principle, provide
means to quantify and integrate relevant information in a comprehensive
framework. The tools developed for addressing these issues are generally aimed
at monetary valuation of impacts and damages and integration of the valued
“external costs” in total cost of the product, e.g. electricity.

The approach of Life Cycle Analysis (LCA) provides a conceptual


framework for a detailed and comprehensive comparative evaluation of energy
supply options. For power generation technologies and transportation fuels, LCA
represents a useful tool to assess the sustainability of different energy
technologies and, by extension, the electric power and transport sectors of
OECD countries.

1. The author would like to thank a number of IEA and NEA colleagues who provided
useful comments on earlier drafts of this note, including Jonathan Pershing, Peter Fraser,
Evelyne Bertel, Laura Cozzi, Lew Fulton and Giorgio Simbolotti.

195
This paper offers a summary of the LCA methodology and an overview of
some of its limitations. It then illustrates, through a few examples, how the
methodology can be used to inform or correct policy making and to orient
investment decisions. Difficulties and issues emerging at various stages in the
application and use of LCA results are discussed, although in such a short note, it
is impossible to address all issues related to LCA. Therefore, as part of the
concluding section, some issues are left open – and areas in which further
analytical work may be needed are described.

2. What is life cycle analysis?

Life cycle analysis (and assessment) is a process that seeks to identify and
assess the environmental, economic and social impacts associated with a
product, process or activity. It does so using various specific but continuously
evolving methodological tools. The assessment covers the entire life cycle of a
product, process or activity from raw material production and transformation to
end use and disposal, in a “cradle to grave” approach. For the case of an energy
product, process or service, life-cycle analysis encompasses all segments
including up-stream and down-stream processes.

Life cycle analysis:

• Identifies the segments of the fuel cycle (or chain) which have the most
harmful environmental impacts (and consequently specific points in the
cycle that provide opportunities for pollution prevention or control).

• Allows for an overall comparison (in a cost and benefit framework) of


short- and long-term economic implications of different energy
technologies or strategies factoring in environmental impacts. An aspect
of this assessment is the valuation of the so-called external costs (and
particularly environmental externalities of the fuel and energy
technology cycles) for policy purposes.

• Identifies research and policy areas for further work.

196
A definition: external costs and externalities

An externality may be defined as:

A cost or benefit that is not included in the market price of a good because it
is not included in the supply price or the demand price. An externality is
produced when the economic activity of one (or a group of) actor(s) has a
positive or negative impact on the welfare function of another actor (or
group of actors) and when the former fails to be fully compensated or to fully
compensate the latter for that impact. Externality is one type of market
failure that causes inefficiency (see Pearce & Turner, 1989).

This definition is most often used in the context of negative environmental


externalities such as air pollution which produce damages to human health,
crops or materials and in which the polluter may not suffer from the direct or
indirect damages. In principle, externalities may also be positive, for
example, the case of a bee-farmer whose bees help pollinate the fruit trees of
a nearby orchard.

Essential to the definition are both the lack of participation in the decision
concerning the economic activity by one or more of the parties affected, and
the absence of full compensation of the costs or benefits accruing to the
receiving party. It should be noted that under this definition, environmental
pollution might conceivably not be an externality if those who suffer from
the negative impacts of that pollution are fully compensated.

3. The life-cycle analysis methodology

Undertaking a life-cycle analysis requires performing the following steps:

• Defining the system’s boundaries.


• Identifying environmental burdens (inventory) and impacts.
• Quantifying and monetising impacts.

The methods for each of these steps are continuously evolving. Although
some of these methodologies have been reviewed and standardised, the methods
and assumptions used by different practitioners still vary widely, as do the
underlying decision criteria for the analysis. As a consequence, the resulting

197
valuations may show differences of an order of magnitude or more. The process
of life-cycle analysis and assessment involves multidisciplinary competencies.
For energy technologies and products these include:

• Engineers and technicians who can define various energy technologies


and their emissions and identify all elements of the fuel cycle.
• Scientists who can identify pollutant dispersion, and who are versed in
issues related to health, ecology, and materials and who can identify and
measure the full chain of environmental impacts.
• Economists to monetise the impacts measured and to develop cost-
benefit assessments.

3.1 Methods and criteria

3.1.1 Definition of the system’s boundaries

The energy system boundaries for the analysis must be clearly defined both in
terms of activities and of their geographical location. To do so, the various stages of
the fuel/technology cycle must be identified. These normally include exploration,
extraction, fuel preparation, transport, conversion, transmission/distribution, use, re-
use and final disposal. The analysis should be extended to plant and infrastructure
construction, operation and dismantling, as well as to waste product management at
all stages of the cycle. A similar path may be developed for technologies that do not
use fossil fuels, like wind-power generation or solar photovoltaic conversion, where
environmental burdens from the processing and use of materials for the plant or
technological device used are more relevant.

In some cases, fuel/energy conversion and use is easily located spatially and
may be the main source of environmental burdens. However, activities upstream
and downstream from actual energy conversion and use are often located
elsewhere. Furthermore, energy commodities and services are normally traded –
often internationally. Fossil fuels used at a given power plant may originate from
many different supply sources. Fuel extraction activities themselves give rise to
flows of wastes and to environmental impacts not necessarily confined to the
immediate vicinity of the activity area. Some pollutants are easily transported by
air and water and can cross national borders, ending up at distant locations: at each
location there may be an effect of the flows of pollutants. Hence the decision
concerning the system and the spatial boundaries set for the analysis is far from
being a trivial one.

198
3.1.2 Identification of burdens and impacts

In principle, a life cycle analysis should strive to identify all environmental


burdens and associated impacts of a specific energy system that fall within the
system boundaries, regardless of whether they are measurable (or even considered
relevant). Scientific investigation improves continuously, increasing our capability
to identify and measure the real environmental impact of polluting emissions; on
occasion, this may lead to a reversal of an early conclusion on the direction of
impacts. Producing comprehensive lists (inventories) of burdens and impacts is
therefore important not just for consistency and transparency, but also because it
allows for an easier updating of the analysis as new information is produced.

This part of the analysis can be performed through the development of an


“accounting framework”, which identifies the various stages of each
fuel/technology cycle and traces the consequences of the associated burdens
through the following stages:
Fuel/technology cycle stage Activity Burden Impacts
This accounting framework is easily presented as a matrix like the one in
Table 1, having in the columns the four terms above, and in the rows, for each
stage of the cycle, the corresponding activities, the burdens produced and the
2
type of impacts. Xs mark indicative priority impacts for further analysis.

The identification of impacts can be done through expert knowledge or


through reference to the available literature on the particular segment of the
system or on the burden considered. This process requires following all the links
of the chain, from the emission of a pollutant (or creation of a burden) all the
way to the receptor(s), and the analysis of the receptor’s response.

Prioritisation of the importance of different impacts (in terms of their


magnitude or of their social, economic or environmental relevance) is then
necessary to isolate those impacts that are worth studying in further detail for
valuation purposes. One example could be materials such as cadmium, used in
the production of photovoltaic systems, whose concentration could lead to
significant health impacts.

2. Detailed accounting frameworks for a number of selected fuel cycles have been
developed within the ExternE project, a collaborative effort launched by the
European Commission and the US Department of Energy to assess the external
costs of fuel cycles. The project was implemented with the collaboration of a large
number of European and US research institutes.

199
Table 1. Schematic example of accounting framework

Impacts
Health Environment
Stages of Activities Burdens Occupational Public Natural Agricultural Man-made
fuel cycle
Fuel Mine Dust X X
extraction construct.
Mining X X X
Waste Dissolved X X
water metals
management
Solid waste Solid waste X
management dumps
….
Fuel Fuel X X
preparat. cleaning

Fuel Construct. X X
transport of facility
Loading
Transport X
Unloading X X
… X
Power Plant X X
generation construct.
Operation Primary air X X X X
pollutants
SOx
NOx X X X X
CO2 X X X X
….
etc… ….. ….

3.1.3 Impact quantification and monetisation

Monetising impacts is definitely the most complex stage of a life cycle


analysis. However, methodologies to quantify and attach a monetary value to the
environmental impacts have been well developed in the scientific and economic
literature; many methods date back more than 40 years. For a proper assessment,
a diverse body of knowledge, which among other disciplines, covers
engineering, material sciences, fluid dynamics, climatology, medical and
biological sciences, ecology, sociology and economics, must be put together in
the impact assessment process. A broadly agreed, standardised methodology for
3
this procedure is the Impact Pathways Methodology. Using this method, the
analyst follows the flow of a pollutant or environmental burden from the point of
emission to the final impact on the receptors affected, and then assigns an
economic value to each impact.

3. The ExternE project has made a significant contribution to the establishment of the
Impact Pathway Methodology.

200
Atmospheric or water dispersion models are used to determine pollutant
transport patterns and concentrations at various locations. For the more reactive
pollutants (for instance, SO2), the formation of other pollutants (e.g. H2SO4) from the
originals is estimated through chemical and photochemical reaction models.
Dispersion models vary from very simple to complex depending on the geographic
range and the number of pollutants and chemical reactions covered. Dispersion
models that simulate transport of inert pollutants are usually reliable, but become
less reliable when they involve chemical reactions besides pollutant transport.

Next, the population or the various receptors exposed to the various


pollutants must be defined and characterised, and the impact estimated through the
use of suitable exposure- (or dose-) response functions. Exposure-response
functions, Y = f (X), where f is the impact function, link exposure to pollutant X
(e.g. concentration of particulate in the air) to specific physical impacts Y (e.g. the
number of cases of respiratory diseases due to concentration of particulate in the
air). As impacts are often negative the function is also called dose-damage
function. These are usually derived from the medical and scientific literature, and
based on epidemiological studies, statistical analysis of field or laboratory data.

The specific form of the impact function depends on the precise


phenomenon being investigated. Impact functions may be linear or non-linear.
Impact functions may have threshold values below which the impact is zero (i.e.
below the threshold, the natural response capability of the receptor – biologic
organism or ecosystem – may prevent or counteract the impact). In a few cases,
the response function has negative values of Y below a positive threshold value
of X, indicating that the pollutant may have a positive or “fertiliser effect” at low
doses. The existence of thresholds has important practical as well as policy
implications. Such levels may be points at which preventative or mitigation
action becomes necessary to avoid health damages or acute crises.

Very few of the known impact functions take into account synergies with other
pollutants or other variables in the biological or ecological system considered,
making it hazardous to transfer results from one specific case to another.

Margins of uncertainty in estimated functions are usually expressed in terms


of “confidence intervals”, by which the true value of the impact is said to lie with
x% confidence within a band of variation centred on the mean estimate.
Estimation of these impacts rests on the scientific solidity of the cause-effect
relationship and on the goodness of fit of the statistical relationship. As noted
above, impact or response functions may or may not include all relevant factors in
the equation; both uncertainty in known variables and the fact that important
variables may not be known are risks inherent in the evaluation process.

201
Once the impact of a measured increase in the environmental concentration
of a pollutant has been quantified, the next step in the assessment process is that
of economic valuation. As with other elements of the life cycle analysis, this
element can range from quite simple to extremely complex.

Some impacts lead to a change in the quantity of a good which is traded


(and which has a market value or price – the case of damages caused by ozone
levels to commercial crops). In such a case, the valuation is straightforward: it is
found by multiplying the unit price by the change in yield. Of course, if the
impact is large enough to affect the market price, this price change has to be
taken into account.

The valuation must include all impacts, whether they affect goods normally
traded in the market or not. In fact, when performing a valuation of
environmental impacts, one typically has to deal with goods that are not readily
traded in the market, but which may still be vital to human welfare, such as
health, the preservation of buildings, monuments, wildlife and bio-diversity, and
amenities such as an unspoiled landscape.

A number of methods have been developed to value these goods. One is the
control cost method based on pollution abatement or mitigation costs. This
approach considers the value of the environmental good to be correctly
approximated by the control or abatement cost sustained to restore it. According
to this approach the question posed is what would be the cost to reduce
drastically certain polluting emissions, to restore human health or a forest
damaged by acid rains, or to clean a building covered by soot. Other useful
indicators are the defensive expenditure consumers choose to sustain in order to
avoid certain negative impacts (e.g. double windows to reduce noise pollution).

Another approach taken by economists in the valuation of environmental


goods such as the ones mentioned, is to try to identify individual preferences
through their Willingness to Pay (WTP) for an environmental good or
Willingness to Accept (WTA) payment in compensation for an environmental
damage suffered. This approach, chosen among others by the ExternE project,
covers a variety of methods and techniques:

• One is the contingent valuation method (CVM), which seeks to identify


consumer preferences through direct questionnaires and interviews,
asking consumers how much they would be willing to pay to avoid a
damage or to accept in compensation for a damage suffered. The same
method can be applied to issues as varied as health risks, water quality
or the aesthetic value of an undisturbed landscape.

202
• Another technique is the hedonic price method, which derives the value
of some environmental characteristics from the WTP for goods in
related markets. A typical case is that of different property values for
homes of equivalent size and characteristics but located in
neighbourhoods with different noise or air pollution levels.

• A third method is that of travel costs, which use expenditures in


recreational activities as a proxy for WTP for natural amenities, parks,
wildlife, etc.

The WTP/WTA approach can be used both for values arising from the
actual use of the environment and for values arising even when no identifiable
use is made of the environment, either at present or in the future. The former are
called use values and can refer both to direct use (clean air in one’s
neighbourhood) or indirect use (e.g. clean air for a cousin living in Bhopal):
altruistic values are in the latter group. Non-use values, on the other hand include
cases such as the interest in preserving an unspoiled Antarctic environment, or
for the conservation of a threatened wildlife species unlikely to be encountered
by any given individual in their lifetime. These latter values are the most
difficult to assess.

3.2 Further issues in life-cycle assessment

3.2.1 Discounting

Regardless of the method chosen, one additional step is likely to be


necessary in the valuation process, due to the fact that the impacts deriving from
the various activities of any fuel cycle stage may take place at a different time
from the activity itself. Costs that take place in the future must be translated,
through the discounting process, into present values. Discounting is particularly
important when the assessment of a fuel cycle or a technology is done for
investment purposes. Discounting is, however, controversial. Part of the
literature (particularly that which starts from a strong environmental perspective)
argues against discounting on ethical grounds, because it places more weight on
the welfare of present generations than on the welfare of future ones. Large
environmental damages taking place 50 or more years from now as a result of
present choice will appear negligible. The bias against future generations
becomes more pronounced with higher discount rates.

Even once the decision has been taken in favour of discounting, the choice
of an appropriate discount rate is not trivial. It is known that discount/interest
rates practised by lending institutions can change substantially depending on the
risk profile presented by the borrower or by the enterprise. Furthermore, market

203
rates (faced by individuals and private investors) differ from public discount
rates (which are usually lower). The latter would appear to be preferable in
assessing environmental goods. But even for governments, discount rates,
inasmuch as they represent the opportunity cost of capital, can vary widely
depending on whether their country is a very poor one with increasing
population or a wealthy one with a stable population. In the current debate,
reasonable discount rates for environmental valuation are believed to be in the
range from 0% to 10%.

3.2.2 Transferability of estimates on environmental values

Life-cycle assessment should, where possible, be site specific, and consider


both the local conditions where an energy activity takes place and where the
environmental impacts are produced (even if these are far apart). Ideally site
specific analyses should also be undertaken for the monetary valuation of
environmental benefits and costs. But it is obvious that doing so ex-novo for
each assessment would be enormously expensive, if not entirely infeasible. It is
therefore becoming more common to apply monetary values from studies made
elsewhere to specific cases and locations.

This transfer can lead to problems. The risk of using response functions from
the literature in contexts different from the ones where the study was made (noted
above) is not the only cause for concern. Problems also arise when monetary
values elicited with contingent valuation methods are applied to a very different
socio-economic context from the original one – leading to significant distortions in
the analysis. Population characteristics such as income, age and lifestyles heavily
influence consumer preferences and are critical elements to be taken into account
when transferring value estimates from one situation to another.

3.2.3 Valuation of health impacts

Economic valuation of health impacts, and particularly increased mortality


impacts, is an especially thorny issue, partly because of the ethical implications
of putting a monetary value on human life. It becomes even more of an issue in
the context of transferability of values.

For occupational accidents (both fatal and non-fatal) in specific activities


pertaining to various stages of the fuel/technology cycle, databases are available
on the compensation paid by companies, based on the characteristics of the
workforce and the typical occupational hazards to which they are subject. For non-
occupational health impacts, a frequently used approach to value mortality impacts
and determine the value of a statistical life (VSL), is WTP for a reduction in risk
of death or WTA for an increase in risk. Through this methodology a value of

204
3.1 M for a “statistical life” was determined based on accidental deaths (from car
accidents). Early studies used to compute the cost of mortality by multiplying this
value times the number of additional deaths caused by a given increase in
pollution. However, many health impacts caused by exposure to pollutants tend to
reduce the life expectancy of the weaker portion of the population (the aged and
the ill) rather than kill adults in the 20-40 years age bracket. The ExternE
methodology has recently opted in favour of the concept of “years of life lost”
(YOLL), i.e. the loss of life expectancy due to risk of death and other health risks,
and values the cost of mortality in proportion to the years of life lost (Rabl and
Spadaro, 2001).

Actual estimates of WTP can be made in three alternative ways:

1) Using the contingent valuation (CV) method where people are asked to
indicate how much they would be willing to pay to reduce by an
hypothetical amount their risk of death from certain familiar activities or
how much they would be willing to accept in compensation for an
increased risk in the same activity.

2) Studying wage differentials (ceteris-paribus) for work activities having


different health risks or occupational hazards.

3) Studying voluntary expenditures on items that reduce death risks from


certain activities (e.g. airbags in cars).

All of the above methods are subject to criticism. CV methods are affected
by differences in personal income levels and actual capability to pay and by age
differences. Actual risks may widely differ from perceived risks in some
activities. The wage-risk method may reflect more conditions prevailing in the
labour market rather than actual preferences on risk levels, and therefore tend to
be lower than the CVM. Furthermore voluntary risks are valued by the
individual much differently from involuntary ones, with the value of the latter
being larger for probabilities of death in the same range. Possible sources of bias
are therefore plentiful and should be kept in mind when transferring VSL
estimates from one country to another or from a probability range to another.

Valuation of morbidity impacts is usually done adding up at least three


main value components:

• The cost of the illness in terms of hospitalisation and treatment


expenses, or mitigation expenses (at market values).

• The value of time lost or foregone earnings (valued at the after tax wage
rate).

205
• The loss of utility due to pain and suffering, usually determined through
CVM methods.

These may or may not cover all the costs (for instance, the value of pain
caused to family or people attending the sick person). However, at least in
developed countries, one can count on a sufficiently extensive literature on
estimates of the value of illness cases.

3.2.4 Valuation of global warming impacts

The impacts of global warming and related external costs are particularly
complicated to measure. Few estimates exists, and these are based on models and
scenarios which cover different time-horizons, use often very different
assumptions on population, income and greenhouse gas emissions growth trends,
and try to simulate changes in global climate patterns from limited understanding
of the underlying drivers. It is from such scenarios that changes in global and
regional climate patterns are translated into impacts on agricultural productivity,
frequency of extreme weather effects, epidemiological trends, human mortality,
etc. The monetised costs of such impacts are then calculated, discounted and
aggregated.

Problematic circularities exist in this process. The timing and intensity of


some climatic impacts depends on the rapidity at which certain concentrations of
GHG are reached in the atmosphere. Damages from climate change are a
function of the capability of a given human population to respond with defensive
or mitigation measures: these are in turn a function of income level. On the other
hand, as the value of climate change is usually assessed based on willingness to
pay principles (also related to income level), the value of the same damage is
higher in richer countries. Issues of intergenerational equity are magnified by
use of discounting of these costs over time horizons of one century or more. As
a result of different choice in the critical parameters, estimates vary often by
several orders of magnitude.

Valuation of the global warming impact of a kg of a GHG, hence, requires


extensive reviews of available literature on estimates of these impacts and
computation of average values. This is the procedure that was followed by the first
ExternE project (EC, 1995a).

In alternative, specific emissions scenarios can be chosen (for example those


of IPCC) and a range of possible impacts computed from there using other

206
models,4 as was done in phase two of the ExternE project (EC, 1999b). In that
project, marginal damages were calculated based on different discount rate options
and for different GHG. The results of the second study indicated marginal values
of carbon between 160 and 170 ECU/tC at 1% discount rate, and between 70 and
74 ECU/tC at 3% discount rate. Values for CO2 were respectively in the range
44–46 ECU/tCO2 and 19-20 ECU/tCO2.

4. How can LCA be used in policy making?

Life cycle analysis can potentially provide important inputs both in policy
making and in private decision making, and especially in the area of energy and
environment policy.

The analysis of impact pathways allows the identification of the segments of


the fuel cycle (or chain) which have the most harmful environmental impacts; the
full LCA also requires the measurement of these impacts. This in turn can give an
indication as to where priority should be placed – or where opportunities may be
most attractive – for technology improvement, as well as for policy intervention.
LCA and the valuation of external costs can also help identify the extent of the
measures needed to reduce the harmful activity.

However, these methodologies and their results must be considered as only


one tool among many. Economic theory has some prescriptive indications about
the optimal level of the polluting activity: it corresponds to the level where
marginal benefits from the activity equals marginal cost (private + external cost).
In perfectly competitive markets, with perfect information and no transaction
costs, this optimum point would be easy to identify. In such a situation, Pareto-
optimality could be restored by simply imposing a Pigouvian tax (per unit of
pollutant emitted) equal to the marginal social damage caused by the pollutant
(Baumol and Oates, 1988).

In an imperfect world, the quantification of the externalities is a big


problem in itself, the optimal tax level cannot be computed and the policy
instruments to correct the problem are bound to be “second-best” solutions. This
does not mean that efforts to quantify external costs should not be made: LCA
and externality valuation attempt to do just that. Neither it means the information
thus collected should not be used in policymaking. But designing cost-effective
policies to deal with externalities in a situation of imperfect information is big

4. Two climate change models, Climate Framework for Uncertainty, Negotiation and
Distribution (FUND) and Open Framework, were used to gain a better understanding
of the issues and obtain estimates of GHG marginal damages. IPCC scenarios IS92a
and IS92d were used as a basis for these assessments.

207
issue in itself. Besides presenting administrative costs of varying magnitude,
policies to control environmental damages, may themselves impose further costs
to society because of bad design. Furthermore, obtaining more information in
order to improve policy design has a cost too. For this reason, policymakers find
themselves always in the situation of balancing this type of considerations and
exploring trade-offs until satisfactory “second best” solutions are found.

There exists a vast economic literature on the relative efficiency of various


policy approaches to address externality problems, and on the merits of price and
market instruments (taxes, subsidies and pollution permit trading) versus
“command and control” instruments (prohibitions, standards, directives, etc.).
Following this debate, governments in the last decade have changed their
environmental policy portfolio towards more market-oriented and decentralised
approaches. The debate, however, is not over, as in many situations insufficient
or asymmetric information, high transaction or policy implementation costs, or
the social hazard of some activities leave little practical room for policy choice.
Although inefficient in many situations and difficult to enforce, mandatory
standards are sometimes the only viable solution. Furthermore, policy making
often has to serve and reconcile multiple objectives, not only health and
environmental protection.

LCA may be particularly helpful in areas where new technology or


developments are underway. In these areas, it can help focus R&D programs in
those segments of the cycle where the technology needs to be improved and
made more environmentally benign. Carrying out the analysis prior to project
development can eliminate costs and prevent damages. A similar use may be
made of LCA in more mature technologies (e.g. nuclear power or fossil fuels);
LCA can point to a variety of opportunities for improving the sustainability of
the full fuel cycle operation, and allow project managers to develop least cost
solutions to manage problems.

LCA can be useful in an investment decision framework, as it offers a more


complete comparison of different energy technology/fuel alternatives that
provide the same energy service, in terms of their short and long-term
environmental impacts. Once these elements are correctly included in the
evaluation of costs and benefits for each technology and accounted for together
with “internal costs”, the investor is provided with a clearer view of the options
available and can make a better choice. Such value is not limited to an investor.
LCA and the dissemination of its results among the general public in a
democratic society can help make informed choices among technological
alternatives; provide input to the political debate on the sustainability of energy
systems; build consensus around specific projects or help reject those that pose
severe environmental threats.

208
The following section of this paper provides a few practical examples to
illustrate these issues, using information on external costs produced through the
application of LCA and externality valuation methods. The first deals with the
transport sector, the second with power generation. Both examples not only
provide further insight into the LCA methodology, but also illustrate how
current policy approaches might be changed were such LCA results to be
factored into investment and decision making.

4.1 Case 1 – Transport

4.1.1 Results from LCA analysis

One of the most thorough evaluations of the LCA approach was that
developed in the ExternE project. In its second phase, the ExternE project
applied the “impact pathway” to the transport sector (Bickel et al., 1997). LCA
was undertaken in a number of case studies in France, Germany, Greece, Italy,
the Netherlands and the United Kingdom. The results provide an overview of the
variation within the European Union, among several site-specific results.

The study covers passengers and freight transport by road, rail and
waterways, but not air transport. Within each type of transport, different
technologies are examined:

• For passenger transport: petrol cars with and without three-way catalysts
(TWC), diesel cars, LPG cars, diesel coaches and buses, various types of
small buses (diesel, LPG, bio-methanol, bio-ethanol, bio-diesel, electric)
electric buses, trams, metros, and trains (electric: local, inter-city, high
speed).

• For freight transport: light goods vehicles (petrol, with and without TWC,
and diesel), heavy goods vehicles (diesel), and trains (electric and diesel).

For passengers transport, travels in different urban agglomerations (Paris,


Athens, Milan), urban areas (Stuttgart, Amsterdam, Barnsley) and extra-urban areas
(motorway drive or inter-city train, Stuttgart-Mannheim) are considered, and the
environmental impacts of each transport option are quantified and monetised.

The analysis starts from quantification of air emissions for each pollutant by
multiplying emission factors by transport distance covered. Emission factors in
turn depend on the technology used (vehicle type), traffic conditions, and driving
cycles. Total emissions are the sum of emissions in each road segment. Then
marginal increases in pollutant concentration are estimated both locally (in a
band 35 km wide on each side of the road) and regionally (the entire European

209
continent, except former Soviet Union) through the use of dispersion models.
Physical impacts on health (mortality and morbidity), crops and materials are
then calculated through exposure-response functions. Finally, economic
valuation of the estimated impacts is performed.

While the authors recognise that substantial uncertainties arise in the


estimate of environmental cost of air pollution, they suggest that the largest
uncertainties lie in the last stage and are particularly related to political and
ethical issues. They also provide uncertainty labels for each impact value, based
on geometric standard deviations (labels A, B, C, in decreasing order of
confidence in the estimate).

The evaluation covers various types of impacts. Health impacts include,


besides mortality, acute and chronic health effects from pollutants emitted in the
transportation activity: particulate, SO2, NOx, CO, ozone, as well as aldehydes,
ethene and ethylene oxide, MTBE (Methyl tertiary-butyl ether), volatile organic
compounds, benzene, butadiene, polycyclic aromatic hydrocarbons, diesel
exhaust particulate and lead. Impacts on the natural (soils and crops) and man-
made environment (materials, buildings), as well as on global warming were
5
also taken into account. Furthermore, in two cases the impacts of the up- and
down-stream processes, based on the detailed life-cycle analyses (fuel
production; production maintenance and disposal of vehicles and infrastructure)
available from ExternE or in the existing literature, were added.

The results for 1995 indicate that health impacts dominate the damages
quantified in that study, in particular mortality due to primary and secondary
particulate, while carcinogens (highly toxic but usually emitted in small
quantities) have a much smaller size. Furthermore, researchers found that the
magnitude of the health impacts, particularly for diesel vehicles, depends on the
population density around a road. This is mostly due to the importance of
primary particles in total damage.

Estimated damages for a diesel car range from 560 mECU/vkm in Paris, to
6
65-107 mECU/vkm in urban areas and 30-38 mECU/vkm in extra-urban areas.
Table 2 gives some additional information on costs.

It is possible to compare these costs against prices experienced by


consumers in 1995 (year the evaluation was performed). For comparison we take

5. Global warming impacts in this study were determined on the basis of IPCC results,
but using two climate damage models. Resulting values, when catastrophic or higher
order damages are neglected, are in the range of those reviewed by the IPCC, i.e. $5-
$125/tC. When higher order damages are included the range of values is greatly
enlarged. Hence the IPCC spread of values was used in the ExternE project.
6. These are the damages stemming from vehicle use only.

210
as reference the 1995 prices (including fuel tax and VAT) for diesel oil and
unleaded gasoline, and average fuel consumption per 100 km, to compute direct
fuel costs in Table 3.

It shows that quantified damage due to airborne pollutants and greenhouse


gases are at least as large as the direct fuel costs for a diesel vehicle, and may be
up to 13 times as large in an urban agglomeration like Paris. On the other hand
for gasoline in a three-way catalyst car, the direct fuel costs are higher than
pollution damages; in fact, damages are more than covered by the fuel tax,
which in the countries considered represents over 70% of the direct fuel cost.

Table 2. Damage estimates for diesel and TWC petrol cars


in different locations given as “best estimate” in mECU/vkm

Impacts Agglomerations Urban areas Extra urban areas


(motorway drive)
Paris Stuttgart Amsterdam Barnsley Stuttgart- Tiel drive
Mannheim
FR DE NL UK DE NL
Diesel car
Particles (PM2.5) 534.090 50.430 78.600 97.400 18.770 29.500
Other primary pollutants a) 4.970 1.663 1.283 2.055 0.781 0.540
Secondary pollutants b) 20.060 10.920 4.900 4.380 8.700 6.100
Global warming 2.970 2.280 2.700 3.450 1.990 2.300
Up- and down-stream c) 7.100 7.100
Total 562.090 72.393 87.483 107.285 37.341 38.440

TWC petrol car


Particles (PM2.5) 53.410 3.730 1.960 4.170 1.100 0.740
Other primary pollutants 1.440 0.320 0.180 1.115 0.094 0.131
Secondary pollutants 18.040 5.210 2.320 4.440 6.600 3.640
Global warming 3.580 2.980 3.200 3.480 2.380 2.490
Up- and down-stream 9.000 9.000
Total 76.470 21.240 7.660 13.205 19.174 7.001

Source: Bickel P., S. Schmid, W. Krewitt, and R, Fredrich (eds.): External Costs of Transport in
ExternE. Publishable report – Contract JOS3-CT95-0004, 1997.
a) SO2, CO, carcinogens.
b) Sulphates, nitrates, ozone.
c) Up- and down-stream costs were only computed for Germany: this estimate could be
transferred to the other country cases, with due caution.

The ExternE report also provides information on other technologies and


modes and shows that vehicles with internal combustion engines have higher
environmental impacts than electric vehicles. Damages per passenger kilometre
for trains range from 1.1 to 6.6 mECU/pkm, with the higher value corresponding
to a diesel train. To these values one must add up- and down-stream costs ranging
from 2 to 5.5 mECU/pkm, mostly occurring in infrastructure production,
maintenance and disposal. The advantage of electric train over road transport for

211
goods is even larger than for passenger transport: 0.8-8.8 mECU/tkm for trains
against 33-400 mECU/tkm for gasoline or diesel vans, with heavy goods road
vehicles posting costs in the middle of the range.

Table 3. Average fuel consumption, fuel price and direct fuel costs for
France, Germany, the Netherlands and the United Kingdom, in 1995

Fuel France Germany Netherlands UK

Average fuel Gasoline 8.5 9.3 8.3 9.2


a)
consumption, L/100km Diesel 6.7 7.6 6.9 7.5

After tax fuel price Gasoline 0.8641 0.8142 0.9156 0.6779


b)
ECU/litre Diesel 0.59 0.598 0.7374 0.655

Direct fuel cost Gasoline 73.4 75.3 76.3 62.4


mECU/vkm Diesel 39.8 45.6 50.9 48.8
a) Source: OECD/IEA (2000): Energy Prices and Taxes, Paris.
b) Gasoline prices are an arithmetic average of two different grade and octane n.
unleaded gasoline. Data have been converted in 1995 ECU.

4.1.2 Existing policies on transport

Governments have adopted a wide variety of approaches to deal with the


environmental and social externalities associated with transport. These range
from imposing standards on automobile emissions to taxing vehicles, to taxing
gasoline. It is interesting to compare the costs associated with the externalities
(as described by the ExternE study above) to those experienced by the consumer.
In the following analysis, we look specifically at taxation.

Historically, the main role of taxes on energy products (particularly oil


products, gas and electricity) has been to raise revenues for governments.
Demand for oil products is rather inelastic with respect to price, which makes
them an excellent candidate for high taxes.

In OECD countries, these taxes represent, on average, slightly less than 6%


7 8
of total tax revenues (6.5% for the EU15) . While definitely less productive than

7. OECD/ENV – Environmentally-Related Tax Database.


http://www.oecd.org/env/policies/taxes/index.htm. Note that the OECD definition
of environmentally-related taxes includes “any compulsory, unrequited payment to
general government levied on tax bases deemed to be of particular environmental
relevance.” This definition includes all energy taxes (which represent around 90%

212
income taxes as a source of fiscal revenues, energy taxes still significant. About
90% of total energy taxes come from motor fuel taxes. Countries dependent on
imported energy have also used taxation as a way to decrease this dependency,
reinforcing the fiscal justification.

In the last decade, a third motivation has appeared with increasing frequency:
the environment. Environmental taxes reflect some of the environmental costs of
using fossil fuels, discouraging the use of the resources taxed. Ideal application
requires that these taxes fully reflect the magnitude of the environmental
externalities associated with the use of the fuel in question.

The definition of environmental externalities used by the OECD implies that


any tax on energy is environmental in nature. A stricter definition, in which taxes
are counted only if levied specifically to help attain environmental goals, sharply
limits the list of environmentally related energy taxes. Under this definition,
environmental taxes on fossil fuels would include CO2 or sulphur emission taxes.
A number of other taxes levied on vehicle ownership based on engine type and
horsepower or on use of highways may also be considered as related to
environmental goals. Currently, however, only 1% of total energy taxes is
environmentally based, in the sense that it is proportional to polluting emissions.

Ultimately, a tax of a given amount per litre or tonne of fuel will have the
same impact on the consumption of that fuel regardless of the specific purpose
9
for which it was created. The impact of a tax depends on the absolute level of
all taxes on a specific fuel, as well as the relative tax on that fuel compared to
the tax on other fuels.

While the purpose of an environmental tax is to discriminate against fuels


that damage the environment or public health, when it is superimposed to a tax
structure designed to promote other objectives, the end result may not support
the environmental goal.

In this sense, transport fuel taxation can present some striking anomalies.
One is the preferential fiscal treatment accorded to automotive diesel fuel in
certain countries. This in part has been supported by the fact that diesel engines
are more energy-efficient than conventional gasoline-powered engines.
However, in many cases (France, Spain, Italy and Belgium) this difference was

of the total) as well as vehicle taxes, but the database also covers fees and charges
for environmental services provided by the government.
8. Commission of the European Communities: Green Paper – Towards a European Strategy
for the Security of Energy Supply. COM(2000)769. Brussels, 29 November 2000, p. 57.
9. Jean-Philippe Barde (OECD/ENV): Taxes environnementales et réformes fiscales
vertes dans les pays de l’OCDE. Paper presented at the Seminar “Les Réformes
Fiscales Vertes en Europe”, Paris, 10-11 October 2000.

213
simply designed to discriminate in favour of commercial road transport,
10
especially of goods (EC: Green paper-2000). Already, diesel vehicles
represented about a quarter of all new cars sold in 2000 in Western Europe. The
preferential tax treatment may in fact have resulted in increased driving as a
11
“rebound effect”, thus erasing much of the potential diesel fuel savings.

Figure 1 on gasoline and diesel oil price and taxes gives an idea of the
variation in tax rates in European countries compared with countries in North
America or the Pacific. Figure 2 shows tax differentials for diesel oil compared
to gasoline.

The ExternE – Transport study can be used to give an indication on the


discrepancy between the current structure and magnitude in transport fuel taxes
in European Union countries and the estimated level of health and environmental
externalities their use produce. To perform this simple calculation we have used
average fuel efficiencies for gasoline and diesel passenger cars in EU countries
in 1995. Then we compared the range of damages with the range of unleaded
gasoline and diesel tax in EU countries in the third quarter of 2001. Results are
shown in Table 4.

While present day gasoline taxes in the EU are high enough to cover the
minimum estimated damages from its use, only in one case (that of the United
Kingdom) do they cover the maximum value. The same is not true for diesel oil.
Taxes on diesel fuel are not only lower than for gasoline (counter-intuitively
with information on health costs), but the highest tax applied in Europe on this
product is only sufficient to cover the lower end of the damage range.

10. Commission of the European Communities: Green Paper – Towards a European Strategy
for the Security of Energy Supply. COM(2000)769. Brussels, 29 November 2000, p. 111.
11. This effect may represent in fact up to half of the fuel savings offered by diesel cars.
See OECD/IEA (2001), Saving Oil and Reducing CO2 Emissions in Transport –
Options and Strategies, Paris (France).

214
Figure 1. Prices and taxes for automotive diesel and gasoline,
3Q2001 (US dollars/litre)
a) Automotive diesel

New Zealand 0.261 0.034


United States 0.27 0.118
Australia 0.222 0.244
Mexico 0.218 0.27
Greece 0.267 0.299
Luxembourg 0.285 0.297
Portugal 0.291 0.301
Spain 0.294 0.321
Czech Republic 0.309 0.321
Poland 0.288 0.344
lovak Republic 0.271 0.37
Ireland 0.324 0.326
France 0.258 0.444
Japan 0.408 0.306
Belgium 0.327 0.392
Finland 0.326 0.396
Netherlands 0.31 0.418
Germany 0.274 0.458
Italy 0.299 0.46
Switzerland 0.31 0.497
Denmark 0.35 0.49
Sweden 0.389 0.462
Austria 0.39 0.513
Norway 0.325 0.589
nited Kingdom 0.29 0.817

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3

b) Unleaded gasoline
United States 0.34 0.101
New Zealand 0.264 0.196
Australia 0.237 0.246
Canada 0.301 0.196
Greece 0.297 0.359
Mexico 0.573 0.086
Luxembourg 0.293 0.399
lovak Republic 0.325 0.369
Spain 0.29 0.423
Czech Republic 0.314 0.406
Ireland 0.312 0.427
Hungary 0.286 0.466
Portugal 0.429 0.369
Poland 0.321 0.49
Switzerland 0.333 0.483
Austria 0.337 0.502
Belgium 0.276 0.594
Turkey 0.34 0.535
France 0.255 0.65
Italy 0.311 0.607
Japan 0.43 0.5
Germany 0.301 0.649
Sweden 0.342 0.624
Netherlands 0.306 0.692
Finland 0.334 0.67
Norway 0.291 0.716
Denmark 0.351 0.669
nited Kingdom 0.289 0.817

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3

Ex-Tax price Tax Component

215
Figure 2. Tax differential between unleaded gasoline and automotive diesel
in IEA countries – Index values based on prices and taxes of 3Q2001*

United States

United Kingdom

Switzerland

Sweden

Spain

Slovak Republic

Portugal

Poland

Norway

New Zealand

Netherlands

Mexico
ountry

Luxembourg

Japan

Italy

Ireland

Greece

Germany

France

Finland

Denmark

Czech Republic

Belgium

Austria

Australia

-0.600 -0.400 -0.200 0.000 0.200 0.400 0.600

Index

Source: The index is computed based on data from OECD/IEA (2001): Energy Prices
and Taxes- Third Quarter 2001. Paris.

*The tax differential index TDI has been computed according to the following formula:
TDI = {[(ETPg + Tg)/ETPg]/ [(ETPd + Td)/ETPd]}–1
where ETPg = ex-tax price of gasoline; ETPd = ex-tax price of diesel fuel;
Tg = tax price on gasoline; Td = tax price on diesel fuel.
Countries showing negative values (to the left of the vertical axis) apply higher overall
(Excise + VAT) tax rates on diesel oil than on gasoline; countries showing positive values
(to the right of the vertical axis) apply higher tax rates on gasoline than on diesel oil.

216
Table 4. Comparison of estimated external costs and fuel taxes for unleaded
gasoline and diesel fuel

Unleaded gasoline Automotive diesel

Average fuel efficiency in


11.331 13.937
km/litre (1995)
Minimum Maximum Minimum Maximum
a)
External cost in Euro-cents/vkm 0.700 7.647 3.734 56.209
External cost in Euro-cents /litre 7.933 86.651 52.043 783.402
External cost in Euro/litre 0.079 0.867 0.520 7.834
b)
Tax range in 3Q2001 in EU-16 0.411 0.935 0.340 0.935
a) Source: P. Bickel, S. Schmid, W. Krewitt, and R. Fredrich (eds.), External Costs of
Transport in ExternE, Publishable report – Contract JOS3-CT95-0004, 1997.
b) Source: OECD/IEA (2001): Energy Prices and Taxes – Third Quarter 2001, Paris
(France).

In fact, to cover the high end of this range, diesel taxes should be in the
neighbourhood of 7.85 Euro/litre. The calculation is rather crude and does not
take into account progress in fuel efficiency and environmental standards made
by passenger cars since 1995. Best fuel efficiencies in Europe (excluding sport
utility vehicles and mini-vans) are currently in the range 4.4-6.6 liters/100 km
12
for diesel cars and 5.8-8.8 for gasoline cars. In many cases product
specifications on the sulphur content of the diesel fuel have been tightened, as
have the standards on abatement of particulate emissions. Newest technologies
allow for a significant abatement level for larger particles (PM10). However,
capturing the smaller particles (PM2.5), which are the most dangerous ones for
human health, is rather more difficult.

These results do not suggest, for example, that a threefold increase in diesel
taxes (to remain in the middle of the external cost range) would be desirable. Not
only would such an extreme taxation level be highly unlikely, but it would also
be an inefficient way to deal with the environmental problems caused by diesel
cars. However, the results of life cycle analysis and logic of the valuation of
external costs do suggest that the fiscal advantage (in fact a true subsidy) granted
to diesel oil should be eliminated, if not reversed, at least until the technological
improvement in diesel engine and pollution-capture systems (especially finer

12. OECD/IEA (2001), Saving Oil and Reducing CO2 Emissions in Transport –
Options and Strategies, Paris (France).

217
particulate) improve to the point that the health impacts are more in line with
those of gasoline cars.

Ideally, one should be taxing particulate emissions by each car, but as


emissions depend (among other issues) on the type of driving and on car
technology, and there is no end-of-pipe way to meter emissions, this is hardly
feasible. Raising and enforcing emission standards for diesel cars or raising
vehicle taxes would be another approach, as would be restricting use of diesel
cars in large urban areas like Paris. In the meantime investment in R&D on
cleaner diesel engines could be increased. But it is clear that something could be
done to strike a better balance between the environment (and our own health)
and our desire for mobility.

A different approach might be required to deal with pollution from diesel


trucks for freight transport. Memories are still fresh of the oil price increase of
2000, which was followed by widespread protests and roadblocks from truck
drivers, and prompted effective tax reductions on transport fuels. Increasing
taxes on diesel fuel for this category of users may need a strong political resolve.
And the strength of the truckers rests on the very simple fact that over 80% (in
some countries over 95%) of all goods in Europe are moved by road transport.
This is a reality that is impossible to change overnight, especially if no measure
is taken to counteract existing trends.

4.2 Case 2: Electric power generation

4.2.1 Results from LCA analysis on cross fuel comparisons

A number of interesting applications of life cycle analysis to power


generation can be found in the recent literature: we shall illustrate a few
examples and then discuss how they can be used in public policy as well as in a
private investment context.

A. Voss at the University of Stuttgart undertook an LCA for a cross-fuel


assessment of electric power generation options in Germany (Voss, 2000).

In that study he used the following seven energy technology options to


represent the current and near future portfolio in German electricity generation:

• Pulverised hard coal fired power plant, equipped with flue gas
desulphurization (FGD), selective catalytic NOx reduction and with a net
capacity of 508 MW and 43% thermal efficiency.

218
• Lignite fired power plant, with FGD and selective NOx reduction, net
capacity 935 MW, and 40.1% efficiency.
• Gas combined-cycle power plant, with 778 MW net capacity and 57.6%
thermal efficiency.
• Pressurised water nuclear reactor, 1 375 MW capacity.
• Photovoltaic home application, with amorphous Silicon cells, 5kW peak
capacity.
• Wind-converter, with 1 MW capacity, 5.5 m/s average wind speed.
• Run-off hydropower plant, with 3.1 MW capacity.

The study, carried out from a sustainability perspective, took into account
some of the key impact categories, including cumulative energy requirements,
raw material requirements, emissions, health risks, external costs and power
generation costs.

The results indicated that cumulative energy requirements, on a kWh of


output basis, for plant construction and decommissioning plus production and
supply of the main fuel used, were highest for the PV installation, followed by
the nuclear plant and then by lignite, coal, wind, gas and finally hydro. A similar
situation was found for materials requirement in plant construction and operation
and in fuel supply: PV had the highest requirements per kWh in such minerals
and metals as bauxite, iron and copper, and was second only to lignite for
limestone requirements. The relatively small energy density of solar radiation
explains this comparatively high material demand. By contrast, on a kWh of
output basis, nuclear or hydro are relatively frugal in their use of materials with
respect to all other technologies.

A slightly different story emerges when total life cycle emissions of CO2,
SO2, NOx and fine particles are considered. The situation is as shown by Table 5.

Emissions from coal and lignite are higher in this case, but the PV still does
not compare favourably with respect to natural gas, for instance, while nuclear,
wind and hydro show similar orders of magnitude in their overall emissions. The
table above, however, although considering some of the pollutants with higher
impacts on human health, does not include other emissions (e.g. in the case of
nuclear, like ionising radiation).

Once all these were taken into account, both in up-stream and in
downstream processes and in plant operation, the total health risks to the
population were estimated, using the impact pathway approach, in terms of years

219
of life lost (YOLL)13 per unit of electric power produced (TWh). Results for the
nuclear fuel chain include the expected value of risks from nuclear accidents
beyond design, rated in this study as negligible.

Table 5. Total life-cycle emissions for


various power generation technologies

CO2 equivalent SO2 NOx Fine particles


(g/kWh) (g/kWh) (g/kWh) (g/kWh)
Coal 951 351 696 64
Lignite 1 072 402 830 263
Nat. Gas CC 410 125 351 37
Nuclear 20 73 48 25
PV 216 433 321 107
Wind 41 68 49 18
Hydro 31 42 45 12
Source: A. Voss, (2000), Sustainable Energy Provision: a Comparative Assessment of
the Various Electricity Supply Options. In Proceedings of SFEN Conference “What
Energy for Tomorrow?”, Hemicycle of the Council of Europe, Strasbourg, 27-
29.11.2000. French Nuclear Society (SFEN), pp. 19-27.

Lignite and coal plants, with 73 and respectively 54 YOLL/TWh, top the
risk scale, mostly due to health impacts from emissions in the plant operation
stage. Third comes the PV plant, with 33 YOLL, entirely due to emissions
occurring in the up- and down-stream processes, followed by the gas CC plant
with about 25 YOLL/ TWh, three-fifths of which due to up- and down-stream
processes. Nuclear compares favourably with about 10 YOLL/TWh, almost
entirely resulting from emissions in the up- and down-stream processes. The best
performers, according to this criterion, are wind and hydro with respectively 4
and 3 YOLL/TWh due to plant operation.

Once these health impacts and other environmental damages due to air
pollution are taken into account and assigned a monetary value, an estimate is
obtained of the external costs not accounted for in the electricity price. Table 6
shows the breakdown of these costs by power generation technology.

For each technology the author of the study then adds these costs to those
for power generation (or private costs), sustained to pay for construction

13. Years of life lost represent the loss of life expectancy due to risks of death and other
health risks.

220
materials, plant and equipment, fuels, and operations and management costs.
The result gives an estimate of the overall social costs of electricity produced
with different technologies and some information on the weight of external costs
in the total (generation + external) cost of each technology. Knowledge of the
magnitude of these two elements is extremely valuable when choosing among
generation options.

Table 6. Quantifiable external costs of energy systems (in Euro-cent/kWh)

Impact Coal Lignite Gas CC Nuclear PV Wind Hydro


Health effects 0.8 1.0 0.3 0.2 0.4 0.05 0.04
Crop losses -0.03 -0.03 -0.01 0.0008 -0.003 0.0005 0.0004
Material damage 0.02 0.02 0.007 0.002 0.01 0.001 0.0007
Noise nuisance 0.006
Acidification/ 0.2 0.8 0.04 0 0.04 0 0
Eutrophicationa)
Global warmingb) 1.6 2 0.8 0.03 0.3 0.03 0.03
Sub-total 2.6 3.8 1.1 0.2 0.8 0.09 0.07

Source: A. Voss, (2000), Sustainable Energy Provision: a Comparative Assessment of the


Various Electricity Supply Options. In Proceedings of SFEN Conference “What Energy for
Tomorrow?”, Hemicycle of the Council of Europe, Strasbourg, 27-29.11.2000. French Nuclear
Society (SFEN), pp. 19-27.
a) Valuation based on marginal abatement costs required to achieve the EU “50% – Gap Closure”
target to reduce acidification in Europe.
b) Valuation based on marginal CO2-abatement costs required to reduce CO2 emissions in
Germany by 25% in 2010 (19 Euro/t CO2).

For the specific case examined, the sum of maximum14 generation costs and
external costs equals:
• 60 -cents/kWh for solar photo-voltaic power.
• 10 -cents/kWh for hydro-power.
• 7.5 -cents/kWh for lignite.
• 7 -cents/kWh for wind-power.
• 6 -cents/kWh for coal.
• 5 -cents/kWh for gas CC.
• 3 -cents/kWh for nuclear power.

14. A. Voss indicates minimum and maximum values for each technology generation
costs.

221
It is important to underline that external costs represent a significant portion
of the total cost for fossil fuel power generation (from about 1/3 for gas CC to ½
for lignite), and a very low or trivial portion of the costs for hydro, wind and
nuclear power. On the other hand, while external costs are non-trivial in absolute
terms for PV, they almost disappear compared to the high generation cost of this
new technology.

A paper by Rabl and Spadaro (Rabl and Spadaro, 2001), reports some
figures on the external costs of various fossil, nuclear and renewable power
generation cycles, computed in the framework of the ExternE project and with
an impact pathways methodology. The results are reported in the Figure 3. In
both analyses, fossil fuel based generation presents higher external costs than
generation from renewables. This is mainly due to the costs of global warming
effects and of health impacts.

Rabl and Spadaro warn the reader about the variability and uncertainties
connected to these estimates, which represent average values. For fossil fuels
and nuclear plants, estimates correspond to typical existing plants in France and
Europe or to new planned ones responding to EU environmental standards. The
nuclear fuel cycle also features small external costs, although it is not so clear
15
whether all the significant impacts of this fuel cycle have been duly quantified.

For renewables, estimates represent averages for the case studies analysed
in Europe within the ExternE project. The range of variability, however, is large
enough that estimates could vary by a factor of four. And in fact the differences
with respect to the situation depicted in Table 6 by Voss are rather significant.
Figures for fossil-fuelled plants in the German example seem to correspond to
those for the newest planned installations considered by Rabl and Spadaro.

Further estimates on the external costs from newer power generation


technologies have been produced by the ExternE project (EC, 1999c). Fuel cells,
for example, present non-trivial values for external costs, especially when they
are fuelled with natural gas.

15. See the Externe website: http://externe.jrc.es/All-EU+Conclusions.htm.

222
Figure 3. Cost of environmental damages for
various power generation cycles. cents/kWh

PV
Other pollutants
Wind CO2
Biomass

Hydro*)
Technologies

Nuclear, 1990

Gas CC >2000

Fuel oil >2000


Fuel oil,1995

Coal >2000

Coal, 1995

0 2 4 6 8 10 12 14 16
cents/kWh
*) Due to the extreme variability of costs depending on the site, the value reported corresponds
to the highest value of the estimate, ranging between 0.04 and 0.74 -cents/kWh.
Source: A. Rabl, J. Spadaro (2001): “Les coûts externes de l’électricité”, Revue de
l’Énergie, No. 525, Mars-Avril, pp. 151-163.

Results from the ExternE study on two CHP plants with a phosphoric acid
(PAFC) and respectively a molten carbonate fuel cell (MCFC), indicate the
following costs:

• For the PAFC system: 0.07-0.13 -cents/MJ heat produced and


1.6-3 -cents/kWh electricity produced.
• For the MCFC system: 0.05-0.1 -cents/MJ heat produced and
16
1.1-2.3 -cents/kWh electricity produced.

This is due to the fact that, while in operation SO2, NOx and particulate
emissions from fuel cells are very small or zero, the production of the plant itself
and the gas fuel cycle contribute considerably to SO2, NOx and particulate
emissions and thus to increased mortality. The production of the platinum and

16. See p. 174, European Commission, DGXII-Science, Research and Development


(1999), ExternE, Externalities of Energy. Vol. 9:Fuel Cycles for Emerging and
End-Use Technologies, Transport and Waste, EUR 18887. Office for Official
Publications of the European Communities, L-2985 Luxembourg.

223
nickel used in the fuel cell stack are very energy intensive. Nontrivial also is the
global warming damage caused by this technology.

4.3 How LCA results can be used

4.3.1 Subsidies for renewables?

A press release of the European Commission – DG Research issued on


20 July 2001, reports on some recent results of the ExternE project concerning
the external costs of electricity production in Europe. Estimates have been
performed according to the impact pathway methodology. This study suggests
that aggregated estimated external costs from electricity generation, according to
the report, would amount to 1-2% of the European Union’s Gross Domestic
17
Product (GDP), not including the cost of global warming.

Table 7 summarises the findings for nine fuel/technology groups for power
generation in 15 EU countries. As shown, wind, hydro and nuclear power present
the lowest external costs, below 1 -cents/kWh, followed by biomass. Fossil fuels
present different ranges of costs: gas is between 1 and 2 -cents/kWh, while oil,
peat, coal and lignite are more in the range 3-10 -cents/kWh. These figures
should be compared with an average power generation cost around
4 -cents/kWh.

To take into account the costs of environmental and health damage, the
European Union report mentions two possible types of policy:
1) Factor the external cost of damaging fuels and technologies in their
electricity prices, via a tax.
2) Encourage or subsidise cleaner technologies that cause much lower
environmental costs.

Considering the enormous difficulty of introducing taxation on an EU level,


the European Commission is opting for the second solution. Community
guidelines on state aid for environmental protection, published by the EC in
February of 2001 in fact foresee that “Member States may grant operating aid to
new plants producing renewable energy that will be calculated on the basis of the
18
external costs avoided”. Based on that principle, the maximum amount of aid
granted to renewable energy technologies must not exceed 5 -cents/kWh.

17. Preliminary work within ExternE has shown that the aggregated external costs of road
transport would amount to another 1-2% of the combined GDP of the EU.
18. “Community guidelines on State aid for environmental protection” (2001/C 37/03).
E.3.3.3, Option 3, §63, Official Journal of the European Communities of 3.2.2001.

224
Table 7. External costs for electricity production
in the EU in Euro-cent/kWh (ranges)*

Country Coal & lignite Peat Oil Gas Nuclear Biomass Hydro PV Wind
Austria 1.1-2.6 2.4-2.5 0.1
Belgium 3.7-15.0 1.1-2.2 0.4
Germany 3.0-5.5 5.1-7.8 1.2-2.3 0.4-0.7 2.8-2.9 0.1-0.3 0.05
Denmark 3.5-6.5 1.5-3.0 1.2-1.4 0.1-0.2
Spain 4.8-7.7 1.1-2.2 2.9-5.2 0.2
Finland 2.0-4.4 2.3-5.1 0.8-1.1
France 6.9-9.9 8.4-10.9 2.4-3.5 0.3 0.6-0.7 0.6
Greece 4.6-8.4 2.6-4.8 0.7-1.3 0.1-0.8 0.5 0.25
Ireland 5.9-8.4 3.3-3.8
Italy 3.4-5.6 1.5-2.7 0.3
Netherlands 2.8-4.2 0.5-1.9 0.7 0.4-0.5
Norway 0.8-1.9 0.2 0.2 0-0.25
Portugal 4.2-6.7 0.8-2.1 1.4-1.8 0.03
Sweden 1.8-4.2 0.3 0.04-0.7
United Kingdom 4.2-6.7 2.9-4.7 1.1-2.2 0.25 0.5-0.6 0.15

* Sub-total of quantifiable externalities.


Source: European Commission, DGXII-Science, Research and Development (1999): ExternE,
Externalities of Energy. Vol. 10: National Implementation. EUR 18528, Office for Official
Publications of the European Communities, L-2985-Luxembourg, p. 6.

However, another critical element to be taken into account in generation


portfolio choice, besides the variability of estimates concerning externalities, is the
spread of generation cost estimates found for some technologies, particularly for
those not yet well established in the market, as is the case for wind and solar PV.
Table 8 illustrates some of the uncertainties in the ranges of costs that may
confront the decision maker when considering investments using different
technologies for centralised power generation. Costs are expressed in -cents per
kWh produced (including investment, fuel, O&M costs) and reflect average costs
for new power plants projects in EU member countries and 2000 fuel prices.

Table 8. Electricity generating costs for


various alternatives – 1990Euro-cents/kWh

Coal & Oil Gas Nuclear Biomass PV Wind


lignite
Minimum 3.2 4.9 2.6 3.4 3.4 51.2 6.7
Maximum 5.0 5.2 3.5 5.9 4.3 85.3 7.2
Source: European Commission – Green Paper: Towards a European Strategy for Energy
Supply (2000).
Note: Production costs are for power generation at 7 000 hours overall availability
(equipment availability for PV) and exclude excise taxes and subsidies.

225
As can be seen, even for some mature technologies (like nuclear, coal and
gas) the range of variation in generation costs is rather large. Costs become a lot
more variable when considering new technologies such as power from biomass,
and solar PV.

The European Commission attitude is very much in line with what


governments in the EU (and in IEA countries) are already doing. Information
collected by the IEA on energy policies and measures implemented in the years
1999 and 2000 to deal with climate change, shows that most recently adopted
fiscal measures are aimed at encouraging technology improvement and diffusion
(OECD/IEA, 2001b). In 2000, more than two-thirds of the fiscal policies
adopted are unambiguous price support measures in favour of renewable energy
or of electric power produced through more environmentally benign
technologies. The emphasis is on fostering the deployment of commercially
available technologies and fuels that have very low greenhouse gas emissions
but are not yet competitive with conventional fuel sources. Direct subsidies for
construction of renewable energy power plants, minimum guaranteed price
schemes for power generation from renewables or tax relief measures are
frequent forms of support. Energy tax reductions or waivers, income-tax
reductions, or tax credits are often granted to power generation from renewable
energy sources.

Green pricing schemes and price support programmes are based on the
consideration that in the near term, GHG reductions will be brought about
through the enhanced use and improvement of already existing technologies
rather than through technologies now at the laboratory stage. Support measures
to speed deployment of existing technologies exploit the reduction of cost that
19
follows increased output – known as the “technology experience curve”.

A correct application of price support mechanisms requires accounting –


and assuming an economic value – for a number of factors. These include the
cost of other technologies/fuel sources; climate and non-climate related external
costs (local air or water pollution, damages to human health as well as to the
natural environment, and to the materials, noise, siting issues, etc.); and the long-
term expected changes in energy activities that could emerge as a result of such
programs. Such factors are likely to differ from country to country, but
assessments like the ones performed under the ExternE project can help set
benchmarks and avoid from the beginning the most obvious distortions.

The information collected by the IEA indicates that new taxes or increases
in existing taxes are the least frequently applied measure. Very few taxes that
have recently been adopted seek to discourage technologies and fuels with high

19. Experience Curves for Energy Technology Policy, OECD/IEA Paris, 2000.

226
CO2 or other pollutant emissions, thus seeking to “internalise” environmental
20
costs. Tax increases are sometimes applied to final consumption of electric
power when it is fossil fuel-based.

The prevailing policy attitude, rewarding environment-friendly technologies


rather than forcing polluting fuels and technologies to bear the full costs of
environmental damages, encounters much less resistance from the public, and
from the fossil fuel lobbies, but does have some potential drawbacks. Subsidies
tend to stay in place beyond their “useful” lifetime, and they can discourage
further, even more desirable technology development.

A more consistent approach, with respect to the results of the ExternE


project would be the elimination of remaining coal subsidies, which still exist,
with different political justifications, in a number of European (France, Spain
and Germany) and IEA countries. The IEA does not consider there to be a
realistic security of supply justification for such assistance to continue

4.3.2 A tool in investment decisions

Life cycle analysis and the valuation of the external costs stemming from a
new energy project certainly adds important information to the process of
assessing different investment options. The potential investor, be it a state
monopoly or a privately owned company, would be well advised to take it into
account in its strategic planning. As an actor operating in a given socio-
economic context, it must pay due attention to public attitudes towards health
and the environment, and follow a code of social responsibility, besides
complying with existing law. In fact a company should be, to the extent possible,
anticipating change in public opinion and legislation, and the best time to do so
is when deciding on a new investment.

Ultimately, however, especially in liberalised and deregulated markets, the


decisive element for a company willing to incorporate environmental concerns in
its decision process is the institutional framework. If a company faces a regulatory
environment (environmental taxes, environmental liability with or without
mandatory insurance schemes, standards, or subsidies for environmentally benign
technologies) which encourages it to pay attention to environmental externalities,
it will do so to the extent possible. Otherwise, a green image with the public will
loose to price competition in the market.

20. The “regulatory energy tax”, implemented by the Dutch government for consumption
of non-renewable energy by small consumers, the “environment tax” imposed by
Denmark on fossil fuel and electricity consumption and the CO2 taxes imposed in
Norway and Sweden are a notable exception.

227
The need to factor LCA in the decision, clarifies some of the
economic/environmental trade-offs although investors must still take risks and
face often large uncertainties and constraints including technological
uncertainty, infrastructure constraints, price variability or changing regulatory
frameworks. Table 8 shown earlier illustrates this problem.

Certainly a utility interested in accommodating a specific growth in power


demand does not usually face the full range of technology options or
uncertainties shown in that table. It has a much more constrained decision
framework when it comes to supply routes for fuels, plant siting options and
land availability, infrastructure availability, environmental regulations, access to
capital financing, etc., which may basically rule out many options. Information
such as the one presented in Table 8 represents average costs, based on average
assumptions on discount rates, plant costs, load or availability factors,
construction lead times, fuel costs and so on. But when evaluating a specific set
of options, the investor may, for example:

• Have to use very different discount rates for a gas combined cycle plant
or for a nuclear one, or for different segments of the same cycle.
• Face a set of delivered prices for the fuel totally different from
international prices.
• Need a specialised plant design (with significantly different investment
costs) due to the specific characteristics of the fuel used.
• Consider different construction lead times due to uncertainties in the
licensing process.
• Consider different sale price hypotheses.

The list could get much longer, but it is clear that all elements of the
investment decision must be carefully evaluated and that sometimes purely
subjective elements, may make the difference between a good investment and a
poor one.

If the external cost estimates (e.g. those of the ExternE project (Table 7) are
added to this already complex picture, generation costs for coal or oil-fuelled
plants could increase by 3 to 15 -cents/kWh. At least in part there would be a
trade-off between investment costs and external costs. As a plant becomes more
complex and uses more sophisticated pollution abatement devices, its cost goes
up, while environmental costs produced during plant operation should be
reduced.

With gas, the range of variability of external costs is smaller than that for
coal and the externalities are also smaller, which could favour this fuel vis à vis

228
coal in future investments. A similar consideration may be made for the nuclear
technology, at present mature enough to have a smaller range of variability in its
production costs, and showing relatively low external costs. Biomass-based
21
generation presents a fairly large range of variability, both for internal costs and
for external ones. On the other hand wind power and PV have fairly low
environmental impacts and external costs, but a large range of variation for
investment and other internal costs.

But again, the ExternE estimates give a range of values, which come from
specific case studies and which should be applied with much caution to different
specific cases: all the caveats in the methodology section about the
transferability of estimates become very important. An in depth assessment of
the life cycle costs (including external costs) of the specific project may yield
rather different figures with respect to the value ranges supplied by ExternE.
This could depend on such elements as siting of the plant (population density,
prevailing climatic conditions), origin of the fuel, infrastructure needs, process
and fuels used in plant manufacturing, etc.

While life cycle analysis and an evaluation of external costs should be


included in the decision making process for a new investment, it is unclear to what
extent an energy company can do it and actually controls the life cycle of the
product it is selling. More likely, a company will try to improve the segments of
the cycle it can control, based on LCA results. This is another reason why using
ready-made parameters and values may not be a good idea, especially for large
projects. Uncertainties in available estimates of externalities could make decisions
much less robust. Considering the range of variation of externality valuation, this
would not impact just decisions on projects with very similar total (private +
external) costs. Hence, an ad-hoc study, using more realistic parameters, may be
safer. On the other hand, for small generation projects, it may be unnecessary or
unjustified to perform a full-scale life cycle analysis. In that case it could make a
lot of sense to use results as the ones available in the literature.

A different situation could be the one in which an investor is facing not just
a generic social responsibility, but also specific policies that impose taxes or
grant subsidies. These policies become a part of the regulatory framework and
must be taken into account even if the costs and benefits they impose do not
reflect the true costs and benefits of the project.

21. See Table 1.3, p. 6 of the volume: European Commission, DGXII-Science, Research
and Development (1999), ExternE, Externalities of Energy. Vol. 10: National
Implementation. EUR 18528. Office for Official Publications of the European
Communities, L-2985 Luxembourg.

229
5. Conclusions and issues for further analysis

In spite of the many limitations and uncertainties underlying life-cycle


analysis and particularly the valuation of external costs, it allows the integration
of environmental impacts and externality considerations in policy making. The
methodology has a wide range of possible applications, covering all sectors of
human activity. Concern for the impact of energy production and use on both
local pollution and global climate makes it a particularly valuable tool in energy
and environment policy.

The possibility LCA and externality valuation methods offer to quantify


those impacts and attach monetary values to them, gives policy makers useful
indicators on which to gauge policies to mitigate environmental and health
damages, or monitor and correct past policies according to rationality and
effectiveness criteria. However, in order to get the prices right, straightforward
application to policy design of LCA and externality valuation results is not
sufficient: policy design needs to become a lot smarter and forward looking and
take into account different factors.

Most examples of LCA cases examined in this paper suggest that the largest
share of total external damages assessed, for most energy technologies studied, are
due to health impacts. Damages due to global climate change, to crop, wildlife or
eco-systems loss, or even to the built environment, are apparently less important.
While this result gives a sound motivation to pollution mitigation efforts, it seems
also to indicate possible sources of bias. Models to predict climate change patterns
and their future impacts on human life and activity still have important limitations
and present enormous uncertainties, as do methodologies to value impacts that
might take place decades into the future. Here, the process of discounting and its rate
could be a critical issue. On the other hand, it has to be acknowledged that these
evaluations necessarily reflect subjective, time-related and species-related values.

This paper has illustrated some examples in which the results of life-cycle
analysis can be (or have been) used to introduce new policies or change their
orientations. However, policymaking encompasses social objectives of which
health, environmental protection and energy supply are only a subset. Within
this subset, LCA can provide useful insights. On the other hand, when criteria
such as income or employment growth, maximisation of tax revenues appear in
the same objective function, as is often the case, LCA cannot be expected alone
to provide the answer.

To date – at least in the examples described above, governments do not


seem to be employing LCA as a deciding factor in policy making. It would,
however, be interesting to know to what extent LCA results are informing
government policies on sustainable development, and to what extent they can be
useful in deriving “sustainability indicators”.

230
Similarly, from the point of view of investors, it is not entirely clear what
impact the results of studies like ExternE are having on investment decisions and
whether they represent useful quantitative tools to help orient their choices
towards more environmentally benign technologies.

Future work might seek to evaluate these questions – both as a way of


refining the LCA methodology, and as a way of evaluating how such methods
might be more usefully applied.

This paper has highlighted some of the risks connected to unqualified use of
some of the results of this methodology and to their transfer across different
social, environmental and technological contexts. It has also noted the
significant variability of some of the estimates of energy cycle externalities.
What is less clear is whether a consensus is emerging, in the LCA community,
on the relative ranking of technologies with respect to the magnitude of their
external costs. This too thus represents a possible area for further work.

It has been noted that ample room exists for further refinement of the
methodology, particularly concerning the swift incorporation of new scientific
results from epidemiological studies, global climate research, and ecosystem
analysis. Continuous progress in the technologies assessed should be also
factored in, especially when using LCA results for policy design. Furthermore,
inasmuch as consumer preferences remain a key element in the economic
valuation of impacts, this methodology should closely reflect and incorporate
change in value systems, both at the societal and at the individual level.

A number of other issues remain open, and would need to be tackled, both
with a solid theoretical framework and with the help of some applied analysis.
One such issue, which we have on purpose avoided, is that of energy supply risk,
its evaluation and the evaluation of security of supply as a positive externality,
with a value attached.

Notwithstanding these caveats – and the consequent need for future work –
the LCA methodology does provide clear value-added in the decision-making
process. It is therefore strongly recommended that such analyses continue to be
developed, and where possible applied to policymaking in the energy sector.

231
BIBLIOGRAPHY

nd
W.J. Baumol, W. Oates, The Theory of Environmental Policy, 2 edition,
Cambridge University Press, 1988.

R. Coase, The Problem of Social Cost, Journal of Law and Economics, 3,


pp. 1-44, 1960.

European Commission, DGXII-Science, Research and Development (1995a),


ExternE, Externalities of Energy, Vol. 1: Summary, EUR 16520 EN, Office for
Official Publications of the European Communities, L–2985, Luxembourg.

European Commission, DGXII-Science, Research and Development (1995b),


ExternE, Externalities of Energy, Vol. 2: Methodology, EUR 16521 EN, Office
for Official Publications of the European Communities, L-2985, Luxembourg.

European Commission, DGXII-Science, Research and Development (1995c),


ExternE, Externalities of Energy, Vol. 3: Coal and Lignite, EUR 16522 EN,
Office for Official Publications of the European Communities, L-2985,
Luxembourg.

European Commission, DGXII-Science, Research and Development (1995d),


ExternE, Externalities of Energy, Vol. 4: Oil and Gas, EUR 16523 EN, Office
for Official Publications of the European Communities, L-2985, Luxembourg.

European Commission, DGXII-Science, Research and Development (1995e),


ExternE, Externalities of Energy, Vol. 5: Nuclear, EUR 16524 EN, Office for
Official Publications of the European Communities, L-2985, Luxembourg.

European Commission, DGXII-Science, Research and Development (1995f),


ExternE, Externalities of Energy, Vol. 6: Wind and Hydro, EUR 16525 EN,
Office for Official Publications of the European Communities, L-2985,
Luxembourg.

European Commission, DGXII-Science, Research and Development (1999a),


ExternE, Externalities of Energy, Vol. 7: Methodology – 1998 Update,

232
EUR 19083. Office for Official Publications of the European Communities,
L–2985, Luxembourg.

European Commission, DGXII-Science, Research and Development (1999b),


ExternE, Externalities of Energy, Vol. 8: Global Warming, EUR 18836, Office
for Official Publications of the European Communities, L-2985, Luxembourg.

European Commission, DGXII-Science, Research and Development (1999c),


ExternE, Externalities of Energy, Vol. 9, Fuel Cycles for Emerging and End-Use
Technologies, Transport and Waste, EUR 18887, Office for Official
Publications of the European Communities, L-2985, Luxembourg.

European Commission, DGXII-Science, Research and Development (1999d),


ExternE, Externalities of Energy, Vol. 10: National Implementation,
EUR 18528, Office for Official Publications of the European Communities,
L-2985, Luxembourg.

European Commission, Green Paper: Towards a European Strategy for the


Security of Energy Supply, COM(2000)769, Brussels (Belgium), November 2000.

European Commission – DG Research (2001), New Research Reveals the Real


Costs of Electricity in Europe, Press release, Brussels (Belgium), 20 July 2001.

http://www.europa.eu.int/comm/research/press/2001/pr2007en.html.

P. Bickel, S. Schmid, W. Krewitt, R. Fredrich (Eds.), External Costs of Transport in


ExternE, Publishable report – Contract JOS3-CT95-0004, 1997.

OECD/IEA (International Energy Agency), Expert Workshop on Life-cycle


Analysis of Energy Systems, Methods and Experience, Proceedings of the
Workshop held in Paris, France, 21-22 May 1992, 1993.

OECD/IEA (International Energy Agency), Comparing Energy Technologies, Paris


(France), 1996.

OECD/IEA (International Energy Agency), Benign Energy: the Environmental


Implications of Renewables, Paris (France), 1998.

OECD/IEA (International Energy Agency), Dealing with Climate Change:


Policies and Measures in IEA Member Countries, Paris (France), 2000.

OECD/IEA (International Energy Agency), (2001a), Energy Prices and Taxes-


Third Quarter 2001, Paris (France).

233
OECD/IEA (International Energy Agency), (2001b), Dealing with Climate
Change: Policies and Measures in IEA Member Countries – 2001 Edition, Paris
(France).
OECD/IEA (International Energy Agency), (2001c), World Energy Outlook –
Assessing Today’s Supplies to Fuel Tomorrow’s Growth, Paris (France).

OECD/IEA (International Energy Agency), (2001d), Saving Oil and Reducing


CO2 Emissions in Transport – Options and Strategies, Paris (France).

D. Pearce, The Valuation of Social Cost, Allen and Unwin, 1978.

D. Pearce, R.K. Turner, Economics of Natural Resources and the Environment,


H. Hempstead, Harvester and Wheatsheaf, 1989.

A. Rabl, J. Spadaro, Les coûts externes de l’électricité, Revue de l’Énergie,


No. 525, Mars-Avril, pp. 151-163, 2001.

T. Tietenberg, Environmental and Natural Resource Economics. 3rd edition,


Glenview, Ill., Scott Foresman, 1992.
A. Voss, (2000), Sustainable Energy Provision: a Comparative Assessment of
the Various Electricity Supply Options. In Proceedings of SFEN Conference
“What Energy for Tomorrow?”, Hemicycle of the Council of Europe,
Strasbourg, 27.-29.11.2000, French Nuclear Society (SFEN), pp. 19-27.

234
LIST OF PARTICIPANTS

AUSTRALIA
Keith Croker
Australian Delegation to the OECD
Louis Wibberley
BHP Sustainable Technology

AUSTRIA
Thomas Augustin
Federal Ministry of Agriculture, Forestry, Environment and Water Management

BELGIUM
David Emmery
ONDRAF
Philippe Opdenacker
ELECTRABEL
Jacques Remacle
Ministère des Affaires économiques

CANADA
David Burpee
Natural Resources Canada
Frans Koch
IEA – Implementing Agreement for Hydropower
Judy Tamm
Atomic Energy of Canada Limited

DENMARK
Ture Hammar
Danish Energy Agency
Michel Schilling
Permanent Delegation, Denmark

235
FINLAND
Pekka Järvinen
Fortum Energy Solutions
Pekka Mäkela
Ministry of Trade and Industry

FRANCE
Didier Beutier
COGEMA
Marc Darras
Gaz de France
Jean-Guy Devezeaux de Lavergne
COGEMA
Jean-Charles Galland
EdF Pôle Industrie – Cap Ampère
Stéphane His
Institut français du pétrôle
Caroline Jorant
Areva/COGEMA
Valérie Moulin
CEA
Marie-Anne Salomon
EdF/R&D Département LNHE
Laure Santoni
EdF/R&D Département LNHE
Caroline Schieber
Nuclear Protection Evaluation Centre (CEPN)
Ari Rabl
École des Mines

GERMANY
Katja Gerling
German Ministry of Economics and Technology – Energy Department

236
Alfred Voss
Stuttgart University

HUNGARY
Zsolt Pataki
Hungarian Delegation to the OECD
József Rónaky
Hungarian Atomic Energy Authority

ITALY
Paolo Frankl
Ecobilancio Italia
Alicia Mignone
Italian Delegation to the OECD
Marcella Pavan
The Regulatory Authority for Electricity and Gas
Renzo Tavoni
ENEA

JAPAN
Kenshi Itaoka
Fuji Research Institute Co.
Hiroshi Kataoka
Permanent Delegation of Japan to the OECD
Hiroyasu Mochizuki
Japan Nuclear Cycle Development Institute (Paris)
Kazuhiro Morimoto
Permanent Delegation of Japan to the OECD
Naohito Okumura
The Energy Conservation Centre

MEXICO
Juan Eibenschutz
Comisión Nacional de Seguridad Nuclear y Salvaguardias

237
NORWAY

Edgar Furuholt
Statoil

PORTUGAL
Maria Arlete de Gouveia
Direccao General de Energia

SPAIN
Pablo Blanc
UNION FENOSA, S.A.
Angel Luis Vivar Rodriguez
UNESA

SWEDEN
Birgit Bodlund
Vattenfall AB
Sven-Olov Ericson
Ministry of Industry, Employment and Communications

SWITZERLAND
Martin Beck
Swiss Federal Office of Energy

UNITED KINGDOM
Jean Cadu
Shell
Paul Freund
IEA Greenhouse Gas R&D Programme
Helen Grimshaw
Department of Trade and Industry, UK
Ron Knapp
World Coal Council

238
David Pearce
University College London

USA
Mark Delucchi
University of California
Elisabeth Lisann
US Delegation to the OECD
John Stamos
US-DOE/NE

INTERNATIONAL ORGANISATIONS
Jean-Marie Bourdaire
World Energy Council
Joseph Spadaro
International Atomic Energy Agency

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT

Kenneth Ruffing

INTERNATIONAL ENERGY AGENCY


Shimon Awerbuch
Ralf Dickel
Laurent Dittrick
Peter Fraser
Lew Fulton
Phil Harrington
Jonathan Pershing
William Ramsay
Maria Rosa Virdis
David Wallace
Clas-Otto Wene

239
NUCLEAR ENERGY AGENCY
Evelyne Bertel
Thierry Dujardin
Luis Echávarri
Stefan Mundigl
Kazuo Shimomura
Peter Wilmer

240
OECD PUBLICATIONS, 2 rue André-Pascal, 75775 PARIS CEDEX 16
Printed in France.